Professional Documents
Culture Documents
The Direct Integration Method: A Best Practice for Relief Valve Sizing....................................................................12
The approach described here is easier to use, and provides more-accurate results, compared to leading valve-sizing methodologies
Process Safety and Functional Safety in Support of Asset Productivity and Integrity.........................................126
Approaches to plant safety continue to evolve based on lessons learned, as well as new automation standards and technology
Sampling Particulate
Materials the Right Way
Remi Trottier and
Shrikant Dhodapkar To obtain a representative sample for particle size
The Dow Chemical Company
characterization, adhere to the golden rules of
I
n the chemical process industries
(CPI) it is often necessary to verify sampling and follow these best practices
material specification at various
points in the process. In that effort,
it is usually impossible — or at the analysis. In this example, we can all sample that has a distribution of par-
very least impractical — to measure appreciate the two main, necessary ticle sizes. This notion that individual
the whole production. Instead, small characteristics of the sample: particles are not identical is referred
samples must be extracted from a 1. That a relatively small sample is to as constitutional heterogeneity
parent population. Such is the case in taken (much smaller than the total (CH). The higher the upper end of the
particle size characterization of bulk available) distribution, the higher the hetero-
solids, process streams and slurries. 2. That the sample be representative geneity. The Gy sampling theory can
While truly representative sam- of the whole (so that the correct di- estimate the variance of this funda-
pling has long been an important goal, agnosis can be made) mental sampling error due to the CH,
a number of current trends are driving Although both points are extremely using Equation (1), [2]:
the incentive for rapid implementation simple concepts, a great deal of dili-
of top-notch sampling strategies to be gence is usually necessary to achieve
the standard, rather than the excep- them. Careless sampling of powders
tion. These trends include the ever-in- or slurries often results in a faulty
creasing demand for superior material conclusion, regardless of whether good (1)
quality in the high-technology indus- analytical techniques are employed. In
tries, more-stringent pharmaceutical that respect, the first item that should Where MS is the mass of the sample,
regulations and higher environmental be considered for a particle-character- ML is the mass of the parent popula-
standards, to name a few. ization study is a sampling protocol tion from which the sample is taken,
Unfortunately many sampling that insures a representative sample ƒ is a shape factor (0.5 for spheres, 1
strategies in use today do not take of the proper size. for cubes, 0.1 for flakes), ρ is the par-
into account the most modern sam- ticle density, cL is the mass fraction of
pling theories (for more on the history Statistics of sampling material in the size class of interest,
of sampling strategies, see box, p. 45), The first necessary step for a good sam- d1 is the average particle diameter
which leads to inaccurate test results pling program is to define the sample in the size class of interest, g is the
and unrealistic material specifications that is needed and clearly specify how granulometric factor [ratio of the di-
that are impossible to verify properly. the sample is taken, including equip- ameter corresponding to the 5th per-
The best practices outlined in this ment specification. It is important to centile of the size distribution to the
article provide guidelines for collect- keep in mind that in particulate ma- diameter corresponding to the 95th
ing representative samples from most terial sampling, the best we can ever percentile of the size distribution (d05/
solids handling and processing equip- achieve is a random sample where d95)], d is the diameter corresponding
ment and then reducing the sample all particles within the parent popu- to the 95th percentile of the distribu-
to the proper size for the analytical lation have an equal chance of being tion (d95). This allows the calculation
technique used in the measurement. sampled, thereby assuming that no of the fundamental error for any size
In addition, an assessment of sam- systematic bias exists in the sampling class in a distribution. If the mass of
pling errors, based on simple statisti- process. Since there is no such thing as the parent population is much greater
cal theories, illustrates the pitfalls of two identical samples, a perfectly ex- than the sample mass, the term 1/ML
sampling methods. tracted sample (random sample) will can be dropped from the equation. A
One of the everyday examples of always be inflicted by a residual error, few important highlights from the
sampling that all of us can relate to called the fundamental error (FE), as above equation:
is when a medical doctor orders blood first postulated by Gy [1]. This is due 1. The variance of the fundamental
to be drawn for routine laboratory to the heterogeneity of any particulate error decreases as the sample size
42 Chemical Engineering www.che.com april 2012
Example of a Sampling problem, with solution
A
fter several customer complaints, an engineer is assigned the responsibility of
setting up a sampling protocol for a ground product that frequently does not d05 100
= = 0.40
meet the specification that no more than 5% of the mass, or volume distribution d95 250
should be greater than 250 microns (Figure 1). This product is sold in lots consisting of
several tons. This specification should be measured at the 99% confidence level. This 5th percentile
product has a density of 2.5 g/mL. Assuming that correct sampling techniques were - 100 µm
used to obtain a random sample, what is the minimum sample size that needs to be 95th percentile
collected and analyzed? -250 µm
Solution: Material
1. Since the mass of the sample is much smaller that the mass of the lot, the equation specification
for the fundamental error estimation [Equation (1)] can be rearranged as follows to < 5% greater
than 250 µm
solve for the minimum sample mass
2. Measure the size distribution on a volume, or mass basis to obtain the diameters Figure 1. Example of size distribution
with information necessary to calculate
corresponding to the 5th and the 95th percentile (Figure 1)
minimum sample mass
3. The 99% confidence level implies that the value of FE is 0.01. The variance of the
2
fundamental error, Var(FE), is 0.01 , or 0.0001. The shape factor (ƒ) can be set at
0.5, assuming that the particles can be approximated by spheres. The particle density (ρ) is 2.5 g/cm3. The fraction of material in
the size class of interest cL is 0.05 (5% > 250 microns). The average diameter in the size class if interest (d1) can be taken as 275
microns (see Figure 1). The granulometric factor (g) is defined as d05/d95 (see Figure 1) is 0.40 for this distribution. Finally, d,
defined as the 95th percentile of the distribution (see Figure 1) is 250 microns. Changing all units to CGS units to obtain the sample
mass (MS) in grams, we obtain the following:
(5)
Please note that not only a sample of 4.8 g is needed, but an analysis technique that can analyze the whole sample needs to be utilized.
❏
increases. Since the variance is curacy and precision. In addition to size, sampling typically becomes less
equal to the square of the funda- sampling error, analytical error will of an issue, and analytical errors take
mental sampling error, the funda- also add to the uncertainty of the over. Table 2 outlines the basic steps
mental sampling error decreases in measurement. With modern particle- for correct sampling.
proportion to the square root of the characterization instrumentation, Grab samples should not be used
sample mass the sampling error will typically be- even if one attempts to mix the bulk
2. The variance of the fundamental come much larger than the expected specimen prior to sampling — for ex-
error is a strong function of the analytical error as the top end of the ample, bulk bags or perhaps a sample
coarse end (95th percentile) of the distribution (95th percentile) exceeds brought to the laboratory. It is simply
size distribution as dictated by the 100 microns. Gy defined each of the not possible to obtain a homogeneous
d3 term. seven error components as an additive mix from blending alone, and there-
The above equation can easily be rear- model where the variance of the total fore such a practice should not be used
ranged to provide the minimum sam- error is as follows: to properly minimize grouping and
ple mass to be used in an analysis. The segregation errors. Pitard [2] showed
TE = FE + GE + CE2 + CE3 + DE +
sample mass estimate is the minimum that the variance of the grouping error
EE + PE (2)
sample size, since additional sources can be compared to the variance of the
of error will contribute to the variance If correct sampling practices are uti- fundamental error as follows:
of the total sampling error. It should lized, the terms GE, CE2, CE3, DE,
be noted that these additional con- EE, and PE are minimized, and are
tributors can be minimized through much smaller that the FE term, for (3)
good sampling practices, and therefore particles sized greater than about
are controllable to a large extent. Gy 100 microns. This minimization of As a rule of thumb, at least 30 sample
broke down the total sampling errors the sampling error can only be accom- increments (N) are recommended to
into seven basic components as listed plished through appropriate selection minimize GE errors.
in Table 1. of sampling equipment for all phases
The mass required to meet a prod- of the sampling and sub-sampling pro- Correct Sampling
uct specification is related to the in- cess. For smaller particle sizes, where Correct sampling implies following
herit degree of heterogeneity in the the heterogeneity of the system de- a few simple rules throughout the
material and the desired level of ac- creases as the third power of particle sampling process as well as using ap-
Chemical Engineering www.che.com april 2012 43
Feature Report
S
ampling became a common, but non-scientific practice first in the mining industry,
then in the pharmaceutical and chemical industries shortly after the industrial revo-
lution. Back in those early days of sampling, although no rigorous theory existed,
scientists and engineers used a common-sense approach based on their intuition and
their experience to guess at the requirements on what constituted a good sample. In
the mid-19th century, Vezin was the first to introduce the concept of a minimum sample
size necessary for obtaining a representative sample, without the benefits of modern
sampling theories. He also invented a sampler that bears his name, and is still in use
today. It was not until the 1950s that the guessing game in sampling was replaced by a
more rigorous discipline, thanks to Gy’s [1] development of the statistical theories behind
sampling. This offered a structured approach to sampling from which all sampling errors
are broken down to basic components.
Materiel properties
Does the material have a broad
size distribution?
Process conditions
Intersystems
Figure 4. Rotary chute sampler Is the process in a pressurized
enclosure?
Sample
cups
Rotation
Drive
axis
Direction
of rotation
Intersystems
Figure 8. The spinning riffler is com-
prised of a ring of containers rotating
under a powder stream
Vp Vp Vp
Vs Vs Vs
Vp = Vs Vp > Vs Vp < Vs
Figure 7. These illustrations of isokenetic sampling from a pipeline show the Figure 9. The chute riffler
sampling velocity (Vs) equal to the process velocity (Vp; left), Vp greater than Vs (mid- splits a sample using a series of
dle), and Vp less than Vs (right) alternate chutes
Slurry sampling dent on such parameters as particle ing fluid velocity (Vp). No sampling bias
The same basic sampling rule where size and density, fluid density and vis- is expected during isokinetic sampling.
all particles have an equal chance of cosity, flowrate and pipe diameter [4]. If the process flow velocity is greater
being sampled must also be followed Turbulent flow, which provides mixing, than the sampling velocity, particle
when sampling from slurries. Knowl- is typically required to keep the slurry inertia causes an excess of larger par-
edge of slurry properties and behavior well mixed before sampling. Pipelines ticles to enter the sampling probe while
is essential to insure proper sampling can be sampled isokinetically using a process flow velocity smaller than
strategies. For instance, sampling a nozzles provided the slurry is well sampling velocity will cause an excess
slurry from a point in a tank, or flowing mixed at the sampling point. Isokinetic of larger particles to avoid the probe.
through a pipeline requires the pres- sampling (Figure 7) occurs when the Therefore, non-isokinetic sampling will
ence of a homogeneous suspension at average fluid velocity in the sampling introduce a bias based on the particle
the point of sampling, which is depen- tube (Vs) is the same as the surround- size distribution.
Chemical Engineering www.che.com april 2012 47
Feature Report
ASTM Standards: ASTM D345 - 02(2010) Standard Test Method for Sampling and
Testing Calcium Chloride for Roads and Structural Applications
ASTM B215 - 10 Standard Practices for Sampling Metal Powders
ASTM D346 / D346M - 11 Standard Practice for Collection and
ASTM C322 - 09 Standard Practice for Sampling Ceramic White- Preparation of Coke Samples for Laboratory Analysis
ware Clays
ASTM D460 - 91(2005) Standard Test Methods for Sampling and
ASTM C50 - 00(2006) Standard Practice for Sampling, Sample Chemical Analysis of Soaps and Soap Products
Preparation, Packaging, and Marking of Lime and Limestone
Products ASTM D75 / D75M - 09 Standard Practice for Sampling Aggre-
gates
ASTM C702 / C702M - 11 Standard Practice for Reducing Sam-
ples of Aggregate to Testing Size ASTM D979 / D979M - 11 Standard Practice for Sampling Bitu-
minous Paving Mixtures
ASTM D140 / D140M - 09 Standard Practice for Sampling Bitu-
minous Materials ASTM E105 - 10 Standard Practice for Probability Sampling Of
Materials
ASTM D1799 - 03a(2008) Standard Practice for Carbon Black—
Sampling Packaged Shipments ASTM E122 - 09e1 Standard Practice for Calculating Sample Size
to Estimate, With Specified Precision, the Average for a Charac-
ASTM D1900 - 06(2011) Standard Practice for Carbon Black teristic of a Lot or Process
Sampling Bulk Shipments
ASTM E141 - 10 Standard Practice for Acceptance of Evidence
ASTM D1900-06(2011) Standard Practice for Carbon Black Sam- Based on the Results of Probability Sampling
pling Bulk Shipments
ASTM D197 - 87(2007) Standard Test Method for Sampling and
Fineness Test of Pulverized Coal International Standards:
ASTM D197 - 87(2007) Standard Test Method for Sampling and BS 3406: Part 1: 1986 British Standard Methods for Determina-
Fineness Test of Pulverized Coal tion of particle size distribution Part 1. Guide to Powder Sam-
pling, British Standards Institute, London (1986).
ASTM D2013 / D2013M - 11 Standard Practice for Preparing
Coal Samples for Analysis ISO/WD: 14888 Sample Splitting of Powders for Particle Size
Characterisation International Organization for Standardization,
ASTM D2234 / D2234M - 10 Standard Practice for Collection of Geneva.
a Gross Sample of Coal
ISO 2859-Statistical Sampling. http://www.iso-9000.
ASTM D2590 / D2590M - 98(2011)e1 Standard Test Method for co.uk/9000qfa9.html, International Organization for Standard-
Sampling Chrysotile Asbestos ization, Geneva (2000).
It is better to sample from a verti- number of cuts (>30) for both the pri- widely used for reducing the amount
cal pipe so that particle segregation by mary and secondary samplers needs of powder to be analyzed to a smaller
gravity can be avoided. In such a situ- to be extracted. Not all situations are representative sample. In this com-
ation, the sampler should be located at alike, and therefore, these samplers mercially available device, a ring of
least ten pipe diameters downstream need to be installed and designed containers rotates under a powder
from any bends or elbows in the pipe. properly to fit the application. flow to be sampled, thereby cutting
Particle diameter has a strong influ- Selection of the proper sampling the powder flow into several small in-
ence on particle segregation by grav- equipment may not always be trivial, crements so that each container con-
ity since the settling velocity is pro- and may depend on material proper- sists of a representative sample. The
portional to the square of the particle ties, type of process, and sample re- spinning riffler is a versatile device
diameter. Gravity starts to play an quirements. Table 3 provides a list of that can handle free-flowing powders,
important role at particle diameters questions to consider when designing dusty powders and cohesive powders.
greater than roughly 50 microns. The a sampling protocol. The operating capacity of this device
best approach, if possible, is to sample varies from 25 mL to 40 L. If only
at the discharge where a cross-stream Sample reduction the small capacity spinning riffler is
sampler (Figure 5) may be used as a Powder sampling is typically done available, the Vezin sampler can be
primary sampler followed by a Vezin at two levels, a gross sample taken used to reduce the gross sample to
sampler cutter to reduce sample size. directly from the process, and then the appropriate quantity suitable for
This allows sampling even in the non- sub-divided into samples suitable for the spinning riffler. The spinning rif-
ideal case where some segregation the laboratory. The spinning riffler, fler, when properly used, is the most
may have occurred in the pipe. A large as illustrated in Figure 8, has been efficient sample divider available.
48 Chemical Engineering www.che.com april 2012
Another commonly used device for crease the number of total increments, Summary
sample reduction of free-flowing pow- it will likely not match the number of Appropriate attention to sampling,
ders is the chute riffler as shown in Fig- increments performed by the spinning sample size reduction and data analy-
ure 9. It consists of alternating chutes riffler. As such, the spinning riffler is sis is the first step towards obtaining
where half of the material discharges the best device for sample reduction reliable analytical results from a batch
on one side and the second half on the and should be used whenever possible. [5]. To obtain a representative sample,
other. The total number of chutes rep- Several standards dealing with pow- one must adhere to the golden rules of
resents the number of increments de- der sampling are available from a num- sampling and follow the best practices
fining the sample. Although the sample ber of organizations. A comprehensive as outlined in this article. ■
can be processed several times to in- list is provided in the box, p. 48. Edited by Rebekkah Marshall
Authors
References Remi Trottier is a research Shrikant V. Dhodapkar is a
1. Gy, Pierre, “Sampling Theory and Sampling scientist in the Solids Pro- fellow in the Dow Elastomers
Practice. Heterogeneity, Sampling Correct- cessing Discipline of Engi- Process R&D Group at The
ness, and Statistical Process Control”, 2nd neering & Process Sciences Dow Chemical Co. (Freeport,
Ed., CRC Press, Boca Raton, 1993. at The Dow Chemical Co. TX 77541; Phone: 979-238-
(Phone: 979-238-2908; Email: 7940; Email: sdhodapkar@dow.
2. Pitard, Francis F., “Pierre Gy’s Sampling The- ratrottier@dow.com). He re- com). He received his B.Tech.
ory and Sampling Practice: Heterogeneity, ceived his Ph.D. in chemical in chemical engineering from
Sampling Correctness, and Statistical Process engineering from Loughbor- I.I.T-Delhi (India) and his
Control”, CRC Press, Boca Raton, 1993. ough University of Technol- M.S.Ch.E. and Ph.D. from the
3. Allen, T., “Particle Size Measurement”, 4th ogy, U.K,, and M.S. and B.S. University of Pittsburgh. Dur-
Ed., Chapman & Hall, London, 1990. degrees in Applied Physics ing the past 20 years, he has
at Laurentian University, Sudbury, Ont. He has published numerous papers on particle technology
4. Turian, R.M., and Yuan, T.F., Flow of Sluries more than 20 years of experience in particle and contributed chapters to several handbooks.
in Pipepines, 3, 1977, AIChE J., Vol. 23, 3, pp. characterization, aerosol science, air filtration He has extensive industrial experience in powder
232–243. and solids processing technology. He has au- characterization, fluidization, pneumatic convey-
thored some 20 papers, has been an instructor ing, silo design, gas-solid separation, mixing, coat-
5. Trottier, Remi and Dhodapkar, Shrikant, and of the course on Particle Characterization at the ing, computer modeling and the design of solids
Wood, Steward, Particle Sizing Across the International Powder & Bulk Solids Conference/ processing plants. He is a member of AIChE and
CPI, Chem. Eng., April 2010, pp. 59–65. Exhibition for the past 15 years. past chair of the Particle Technology Forum.
Circle 22 on p. 82 or go to adlinks.che.com/40266-22
Chemical Engineering www.che.com april 2012 49
BS&B Safety Systems
Feature Report
Engineering Practice
W
Integration Method is easier than
hat if someone were to merical integration of the isentropic ever to use to size relief valves
tell you that there is one nozzle equation is now easier, faster,
method available for sizing and more accurate than other meth-
relief valves that applies ods for determining the mass flux pressure at constant entropy over the
to virtually every situation, includ- through a relief valve. This article dis- pressure range encountered in the
ing two-phase flow and supercritical cusses the use of process simulators nozzle. To solve the integral analyti-
fluids? And what if they told you that to simplify the numerical integration cally, an equation of state needs to be
method is more accurate and easier method, and describes the advantages available for the fluid at constant en-
to use than traditional methods or of numerical integration over other tropy. However, for many fluids, such
formulas? As it turns out, both of methods that may be used to calculate an equation is not available for density
these statements are true. The ap- the required relief valve area. as a function of pressure. To overcome
proach described here — the Direct this limitation, various simplifying as-
Integration Method — involves nu- Calculation methods sumptions were traditionally made to
merical integration of the isentropic Isentropic Converging Nozzle allow the integral to be solved analyti-
nozzle equation [1]. Equation. The calculation of the the- cally, rather than by performing a nu-
From as early as 2005, the “method oretical mass flux for homogeneous merical integration.
of choice” for determining the flow fluids through a relief valve is gener- For instance, for non-flashing liq-
through a relief valve has been the Di- ally accepted to be modeled based on uids, the density is assumed to be con-
rect Integration Method [2]. API 520 the isentropic converging nozzle. The stant, and the integral is easily solved.
has also sanctioned this method due isentropic nozzle equation is devel- The traditional vapor-sizing equation
to its general applicability to any situ- oped from the Bernoulli equation by is obtained by assuming the vapor
ation where the fluid is homogeneous assuming that the flow is adiabatic is an ideal gas with a constant heat
[1]. However, because this method is and frictionless [4]. capacity [5]. However, the assump-
perceived to be difficult or time con- tions required by these methods may
suming, many engineers continue to introduce large errors under some
(1)
opt for older, simplified methods, even conditions. In contrast, the Direct In-
though such methods can produce less- tegration Method has been shown to
accurate results. For instance, without The required nozzle area of the relief produce more-accurate results.
careful analysis, using the traditional valve is calculated using Equation (2). Direct Integration Method. The
gas-phase equation near a fluid’s Direct Integration Method uses a nu-
critical point can yield an undersized merical method to evaluate the inte-
valve [3]. (2) gral in the isentropic nozzle equation
Fortunately, thanks to the wide- [2]. API 520 proposes the use of the
spread availability of process simu- To use Equation (1), the fluid den- Trapezoidal Rule, shown below, to cal-
lators and spreadsheet software, nu- sity must be known as a function of culate the integral:
54 Chemical Engineering www.che.com April 2013
Nomenclature1
G0 Mass flux, Pn Nozzle exit Kd Discharge coefficient,
lb/h • in.2 pressure, psi unitless ing predictions are fairly accurate at
ρ Density, lb/ft3 A Orifice area, in.2 Pi Pressure at stage i, psi conditions well away from the critical
P0 Relieving W Relieving mass rate, ρi Density at stage pressure. However, two-phase models
pressure, psi lb/h i, lb/ft3 are more difficult to implement. Exist-
ing two-phase flow models approxi-
1. Unit conversion may be required, depending on the units selected.
mate the pressure-density relation-
ship of the fluid in order to calculate
mass flux will occur at some pressure the integral in Equation (3).
(3) above the relief-valve exit pressure. One of the simplest models, the
This is called the choked pressure [6] Omega Method, assumes a linear
pressure-density relationship, with
The method is performed by using a Implementation the omega parameter (ω) represent-
process simulator to generate data It is possible to fully automate the ing the slope of the pressure-density
points for the fluid density at various Direct Integration Method using a curve. An analytical solution to the
pressures, utilizing an isentropic flash spreadsheet program (such as Micro- isentropic nozzle equation was devel-
routine over a pressure range from the soft Excel 2010) and a process simula- oped using the omega parameter to
relieving pressure to the exit pressure. tor (such as AspenTech HYSYS 7.2) [7]. solve the integral [8].
The simulation data are used to deter- Users can automate the process to the The TPHEM Method uses three
mine the theoretical mass flux at each point where all they would need to do is pressure-density points to define co-
point. simply hit a button in the spreadsheet efficients for an empirical equation
Using Equation (3), the maximum program and the numerical integration of state “model” [9]. The empirical
mass flux is determined by calculat- will be performed on an existing stream equation is then used to evaluate the
ing the mass flux over incrementally in the simulator using a VBA (Visual integral numerically. Pressure-den-
larger pressure ranges, beginning at Basic for Applications) program. sity data for these models are often
the relieving pressure, and observing First, the spreadsheet is set up to provided by a process simulator. If
where a maximum flux is reached. If accept the pressure and density data a simulator is available, then it
the maximum occurs at the relief-valve for the numerical integration points. is much simpler to use the Direct
exit pressure (built-up backpressure), The inlet and outlet pressure points, Integration Method.
then the flow is not choked. Generally pressure step size, and name of relief The Direct Integration Method is
accurate results can be obtained with stream in the simulator are placed fundamentally different from the
pressure increments as large as 1 psi, into specific cells in the spreadsheet, other methods described here because
but smaller step sizes can be specified which are referenced in the VBA code. it does not generate an explicit equa-
if desired [2]. Once the mass flux is The VBA code instructs the simulator tion-of-state model to relate pressure
determined, the required relief valve to create a new, ideal expander pro- and density. Instead, pressure and
orifice area* can be determined from cess block and associated streams in density data are generated using the
Equation (2). the simulator. The code then iterates full thermodynamic models available
The value of the discharge coefficient, across the pressure range and modi- in the selected process simulator, and
Kd, depends on the phase of the fluid fies the pressure of the expander prod- these data are then used to solve the
and varies by the manufacturer of the uct stream and automatically exports integral numerically. Since there is
relief valve. The discharge coefficient the pressure and density data to the no reliance on a curve-fit pressure-
corrects for the difference between the Excel spreadsheet. density model, the Direct Integration
theoretical flow and the actual flow For each data point in the spread- Method is more exact and reliable,
through the nozzle. This value is deter- sheet, the summand, cumulative sum, assuming the simulator’s thermody-
mined empirically for liquid and vapor and mass flux are calculated using namic model is accurate. Specifically,
and reported by vendors for each make Equation (3) with typical spreadsheet there is no chance for inaccuracies
and model of relief valve. If vendor formulas. When a maximum mass flux associated with the fluid equation of
data are not available, an initial guess is reached, the spreadsheet uses this state “model” propagating through the
of 0.975 for gases, or 0.65 for liquids maximum flux value to calculate an rest of the calculations resulting in
can be used [1]. orifice size, given the relieving mass inaccurate mass flux estimations and
For two-phase flow, the liquid-dis- rate and coefficients. Alternatively, ultimately an inappropriate relief-
charge coefficient should be used if the data can be collected using the valve area [8, 9, 10].
flow in the valve is not choked and the “databook” feature in the simulator Note that the Direct Integration
maximum mass flux will occur at the and copied into the spreadsheet using Method assumes that the two-phase
relief-valve exit pressure. If the flow is a simple copy-and-paste operation. fluid is homogeneous, and that the
choked, then the gas-discharge coeffi- fluid is in mechanical and thermody-
cient should be used and the maximum Two-phase relief scenarios namic phase equilibrium. The homo-
The existing single-phase vapor and geneous assumption is valid for most
* While relief valves are designed with a nozzle, non-flashing liquid methods are rela- two-phase reliefs due to high velocity
the area at the end of the nozzle is commonly
referred to as the “orifice area”. tively easy to calculate and the result- in the nozzle, which promotes mixing
Chemical Engineering www.che.com April 2013 55
GE/Consolidated and Allied Valve
Engineering Practice
In Brief
CPI project lifecycle
Process Hazard
Identification
T
When to use a given
he chemical process industries (CPI) vamps and decommissioning.
method
handle a wide variety of materials, In each of the plant life phases, different
Safe-design options many of which are hazardous by na- choices are made by engineers that have a
addressing hazards ture (for example, flammable, toxic direct impact on the overall risks in the fa-
early or reactive), or are processed at hazardous cility; however, the highest opportunities for
conditions (such as high pressures or tem- cost-effective risk reduction are present in
final remarks
peratures). The risks associated with CPI the earlier phases of the project. In contrast,
facilities not only extend to the plant person- the cost of implementing changes in the later
nel and assets, but can potentially affect the stages of the project increases dramatically.
surrounding population and environment — Hence, it is important for the design team to
sometimes with consequences having re- identify risks, and implement effective design
gional or international scale, as in the case of solutions as early as possible.
toxic vapor or liquid releases. This article covers some of the typical de-
It is for this reason that process safety is cisions that the project design team has to
recognized as a key element throughout the make over the course of a project, with ex-
entire life of the plant, and several industry and amples of how the incorporation of process
professional associations and government safety throughout the entire design process
authorities have issued norms, standards and can significantly reduce the risk introduced
regulations with regards to this subject. by a new CPI facility, while also avoiding po-
Process safety, as defined by the Center tential cost-overruns, or unacceptable risk
for Chemical Process Safety (CCPS), is “a scenarios at later stages.
discipline that focuses on the prevention and
mitigation of fires, explosions and acciden- CPI project lifecycle
tal chemical releases at process facilities. A project for a new chemical process facility
Excludes classic worker health and safety usually involves different phases, which are
issues involving working surfaces, ladders, outlined here:
protective equipment and so on.” [1] Process A screening or visualization phase. In this
safety involves the entire plant lifecycle: from phase, the business need for the plant is
visualization and concept, through basic assessed. Typical choices at this stage in-
and detailed engineering design, construc- volve defining plant throughput, processing
tion, commissioning, startup, operations, re- technology, main blocks and plant location
50 Chemical Engineering www.chemengonline.com august 2015
Project Life Figure 1. The relative influence of decisions on
total life cost, and cost of implementing changes
throughout the project lifecycle
Influence of
earlier in the project lifecycle have the
decisions on greatest impact on the total plant life
total life cost
cost; in contrast, the cost of imple-
menting changes in the later stages
Cost of of the project increases dramatically,
implementing
changes
as can be seen on Figure 1.
The same holds true for overall
plant risk, as the impact of decisions
on overall facility risk is greatest in
the earliest stages of the project.
mability, toxicity, reactivity and so with the operating conditions material embrittlement
on); these properties cannot be (pressure, temperature), and ma- ❍❍ higher material inventories
separated from the chemicals terial inventories. As general rules: increase the impact of poten-
• Extrinsic — dependent on de- ❍❍ higher pressures increase the tial releases, whereas lower
sign decisions. As an example: impact of potential releases, material inventories reduce
hazards associated with heating whereas vacuum pressures response times in abnormal
flammable materials with direct increase the probability of air operating conditions
burners can be avoided by using entering the system • Equipment failures — associated
indirect heating ❍❍ higher temperatures increase with damages to plant equipment
the energy of the system (and • Utility failures — associated with
2. Their source: hazards, especially when failures in utilities supplied to the
• Process chemistry — associated near the flashpoint or self-ig- facility, such as electricity, cooling
with the chemical nature of the nition temperature), whereas water, compressed air, steam, fuel
materials (for example, flammabil- very low temperatures could or others
ity, toxicity, reactivity and
Mu?llerGmbh_Chemical so on)
Engineering e pose the risks of freezing,
86x123_2011.qxd:Mu�ll • Human activity — associated with
• Process variables — associated Chem eng 1-4pgKrytox Ad 11-9-2014.qxp_Layout 2 11/15/14 12:46 PM Page
Chemical Engineering e UC 86x123 formation
03/2011of hydrates, or activities by humans over the facil-
s
– Compliant with FDA and cGMP
guidelines
PROTECT PUMPS
DRY RUNNING • CAVITATION • BEARING FAILURE • OVERLOAD
PUMPING
POWER
AMPS
Circle 18 on p. 74 or go to adlinks.chemengonline.com/56200-18
Some studies have more impact HAZOP can serve to detect the need www.processsystems.sandvik.com
in some phases than in others. For for additional safeguards that were
Circle 26 on p. 74 or go to adlinks.chemengonline.com/56200-26
FeatuCompilati
o identify hazards, and address them
neering from Universidad de Buenos Aires (Argentina,
1996). She has published articles and delivered world-
2014 using the best engineering practices wide seminars focused in the use of simulation tools for
the process industry.
in safe-design from early on in the
Alberto Baumeister is the
project lifecycle, along with prop- CEO and co-owner of the Ecotek
erly timed and executed PHAs, can group of companies (same address
be very valuable in avoiding costly as above; Email: abaumeister@
ecotekgrp.com). He has experience
changes during the EPC phase, or in corporate management, project
even worse: potential damages to management, and senior process
persons and the environment. n consulting in engineering projects
for the chemical, petrochemical,
Edited by Gerald Ondrey petroleum-refining, oil-and-gas,
25139
electrical power-generation and agro-industrial indus-
tries. He has a specialization in environmental engineer-
References ing (gas effluents treatment) from the Universidad Miguel
1. Center for Chemical Process Safety (CCPS), “Guidelines de Cervantes (Spain, 2013), a master’s diploma in water
Find this and other related for Investigating Chemical Process Incidents,” 2nd edi- treatment management from Universidad de León (Spain,
tion, CCPS, AIChE, New York, N.Y., 2003. 2011), a specialization in management for engineers at
reference material at Instituto de Estudios Superiores de Administración (Ven-
2. CCPS, “Guidelines for Engineering Design for Process
store.chemengonline.com Safety,” 2nd ed., CCPS, AIChE, New York, N.Y., 2012. ezuela, 1990), and a degree in chemical engineering
from Universidad Metropolitana (Venezuela, 1987). He
3. AACE International Recommended Practice No. 18R- was professor of the Chemical Engineering School at
97, Cost Estimate Classification System – As Applied Universidad Metropolitana between 1995 and 2007, and
in Engineering, Procurement, and Construction for the has written a number of technical publications.
Process Industries.
58 Chemical Engineering www.chemengonline.com august 2015
Feature Report
IN BRIEF
RISK SOURCES FOR SIS
MAINTAINING LOW
FAILURE RISK
EXTENDING PROOF-TEST
INTERVALS
TRACEABLE
CALIBRATION
VERIFICATION
REDUNDANT
REFERENCES
LIFECYCLE
MANAGEMENT TOOLS
S
DETECTING PROBLEMS
uccessful implementation and man- FIGURE 1. Flowmeters like the one shown here can play key
CONCLUDING REMARKS roles in reducing risks with safety instrumented systems
agement of a safety instrumented (SIS)
system (SIS) requires designers and
operators to address a range of ISA 84 (International Electrotechnical Com-
risks. First among these involves the specifi- mission; Geneva, Switzerland; www.iec.ch
cation of a proven measurement instrument, and International Society for Automation;
such as a flowmeter (Figure 1), and its proper Research Triangle Park, N.C.; www.isa.org),
installation for a given application, an under- which provides “good engineering practice”
taking that is fundamental to achieving the guidance for SIS development and manage-
initial targeted risk reduction. ment. The emerging IEC 61511 Edition 2 in-
Second is the definition of the support re- troduces some changes to these guidelines,
quired to keep the flowmeter (or other mea- strengthening emphasis on the requirements
surement subsystem) available at that tar- for end users to collect reliability data to
geted level of risk reduction throughout the qualify or justify specifications and designs.
life of the SIS equipment. The support for the This article shows how to address those
flowmeter must be defined in the design and risks and describes several tools, capabilities
implementation phase. and procedures that can be considered for
Third involves following the recommen- designing and managing a SIS installation in
dations found in the standard IEC 61511/ flow-measurement applications.
50 CHEMICAL ENGINEERING WWW.CHEMENGONLINE.COM AUGUST 2016
Risk sources for SIS 3.0E-02
Under IEC 61511-ANSI/ISA 84, op-
erators and SIS designers are re- 2.5E-02
quired to qualify the appropriateness
tem may not only have high diagnos- FIGURE 4. The diagram illustrates the relationship among the various subsystem elements of a flowmeter
I NNOVATION
Low Liquid Rate Technology
Engineered for demanding, very low liquid flow distillation
Many applications, such as vacuum distillation, fatty acid, tall oil, vitamin production and
natural gas dehydration, operate at very low liquid rates. Spreading the liquid evenly
across the tower, which is difficult to do at low liquid rates, is important for the
good liquid distribution needed to achieve superior packing efficiency.
United States (316) 828-5110 | Canada (905) 852-3381 | Italy +39-039-6386010 | Singapore +65-6831-6500
For a complete list of our offices, visit our Web site.
www.koch-glitsch.com
For related trademark information, visit http://www.koch-glitsch.com/trademarks.
This technology is protected by one or more patents in the USA. Other foreign patents may be relevant.
Circle 29 on p. 82 or go to adlinks.chemengonline.com/61498-29
Circle 14 on p. 82 or go to adlinks.chemengonline.com/61498-14
Column Instrumentation
Basics An understanding of instrumentation is valuable in
evaluating and troubleshooting column performance
Ruth R. Sands
DuPont Engineering
Research & Technology
I
nstrumentation is critical to un- '5
derstanding and troubleshooting
all processes. Very few engineers '5
specialize in this field, and many
learn about instrumentation through
'5
experience, myth and rumor. A good
understanding of the various types of &SSPS
instrumentation used on columns is a
valuable tool for engineers when evalu- '5
ating column performance, starting up
new towers or troubleshooting any type
of problem. This article gives an over-
view of the common types of instru- Figure 1. Which flowmeter is the most Figure 2. Flush-mounted diaphragm
accurate? What is the source of error in pressure transmitters are common in
ments used for pressure, differential the material balance? low-temperature services
pressure, level, temperature and flow.
A discussion of their accuracy, common 1. Based on the material balance, the as in scrubbers and storage tanks.
installation problems and troubleshoot- engineer concluded that the bottoms The process diaphragm, an integral
ing examples are also included. flowrate must be in error and wrote part of the transmitter, is mounted
The purpose of this article is to pro- a work order to have the flowmeter on a nozzle directly on the vessel, and
vide some basic information regarding recalibrated. The instrument group the transmitter is mounted directly
the common types of instrumentation disagreed heartily. By the end of this on the nozzle.
found on distillation towers so that article, the reader will understand the
process engineers and designers can instrument group’s response. Remote-seal diaphragm
do their jobs more effectively. Used in higher temperature service
Pressure when the electronics must be mounted
Introduction There are three common types of pres- away from the process, a flush-mounted
Anyone trying to complete a simple sure transmitters: flush-mounted diaphragm is installed on a nozzle at
mass balance around a column under- diaphragm transmitters, remote-seal the process vessel. A capillary tube
stands that process data contain some diaphragm transmitters and impulse- filled with hydraulic fluid connects the
error. Closing a mass balance within line transmitters. All use a flexible flush-mounted diaphragm to a second
10% using plant data is usually consid- disk, or diaphragm, as the measuring diaphragm, which is located at the re-
ered very good. Generally, some values element. The deflection of the flexible motely mounted pressure transmitter.
must be thrown out when matching a disk is measured to infer pressure. The hydraulic fluid must be appropri-
model to plant data. Understanding The diaphragm can be made of many ate for the process temperature and
which measured plant data is likely different materials of construction, pressure. Hydraulic fluid leaks will
to be most accurate is invaluable in but the disk is thin and there is little lead to errors in measurement. Cali-
making good decisions about a model tolerance for corrosion. Coating of the bration is complex because the head
of the plant, column performance and diaphragm leads to error in the mea- from the hydraulic fluid must be con-
future designs. surement. The instrument accuracy sidered. The calibration changes if the
The following is a real case and a of all three types of pressure trans- transmitter is moved, the relative po-
telling example of how little the aver- mitters is similar, usually 0.1% of the sition of the diaphragms changes or if
age chemical engineer may understand span, or calibrated range. the hydraulic fluid is changed.
about instrumentation. A process engi-
neer with over 20 years of experience Flush-mounted diaphragms Impulse-line
was doing a material balance around a These pressure transmitters are com- Impulse-line pressure transmitters
distillation tower, illustrated in Figure mon in low-temperature services, such can either be purged or non-purged.
48 Chemical Engineering www.che.com March 2008
Purged impulse-line pressure trans- different type of pressure transmitter. strument mechanics. The operating
mitters measure purge-fluid pressure Adequate freeze protection on the im- range must be clearly defined; for ex-
to infer the process pressure. Most pulse lines is also important to obtain ample, is the range 100-mm Hg vac-
commonly, the purge fluid is nitro- accurate measurements. uum, 100-mm Hg absolute, or 650-mm
gen, but it can also be air or other Example 1. A good example of a prob- Hg absolute? Using different measure-
clean fluids. The purge fluid is added lem with impulse-line pressure trans- ment scales in the same plant is con-
to an impulse line of tubing to detect mitters can be found in Kister’s Distil- fusing, and it can make it very hard
pressure at the desired point in the lation Troubleshooting [2]. Case Study for mechanics to calibrate the pressure
process. The purge fluid enters the 25.3 (p. 354), contributed by Dave transmitters accurately.
process and must be compatible with Simpson of Koch-Glitsch U.K., de- Another issue is measuring the
it. Check valves are required to en- scribes three redundant impulse-line relief pressure. The system designer
sure that process material does not pressure transmitters used to mea- must consider the instrument ranges
back up into the purge-fluid header. sure column head pressure. Following available and the accuracy of the
The system must be designed so that a tray retrofit, operating difficulties measurement for the operating range
the pressure drop through the im- eventually led to suspicion of the head versus the relief pressure range. It is
pulse line is negligible. A pressure pressure readings. The impulse lines good practice to install a second pres-
transmitter measures the purge-fluid and pressure transmitters had been sure transmitter on vacuum towers to
pressure with a diaphragm to infer moved during the turnaround. The measure the relief pressure.
the process pressure. transmitters had been moved below Example 2. An excellent example of
the pressure taps on the vessel. Con- calibration problems is illustrated in
Non-purged, impulse-line densate filled the impulse lines and vacuum service in Reference [2]. Case
Rather than a purge fluid, this type caused a false high reading. Relocat- Study 25.1 (p. 348), contributed by
of pressure transmitter uses process ing the transmitters to the original Dr. G. X. Chen of Fractionation Re-
fluid. Usually, this style is chosen location above the nozzles solved the search, Inc., describes several years
when the process is non-fouling or problem by allowing condensate to of troubleshooting a steam-jet system
it is undesirable to add inerts to the drain back into the tower. in an attempt to achieve 16-mm Hg
process. One example is a situation absolute head pressure on a tower. It
where emissions from an overhead Transmitters in vacuum service was eventually determined that the
condenser vent must be minimized. Pressure transmitters in vacuum ser- calibration of the top pressure trans-
An impulse line is connected from vice are generally the most problem- mitter was wrong, and they had been
the desired measurement point in atic, leading to greater inaccuracy in pulling deeper vacuum than they
the process to a pressure transmitter, the measured value. Damage to the thought. The top pressure transmitter
which measures the process pressure diaphragm can occur from exceeding was calibrated using the local airport
at the remote point. The system must the maximum pressure rating of the barometric pressure, which was nor-
be designed so that the pressure drop instrument. Often, this happens on malized to sea-level pressure and was
through the impulse line is negligible. startup, or it can happen when per- off by 28-mm Hg.
The system designer must consider forming a pressure test of the vessel.
the safety implications of an impulse- The diaphragm deflects permanently Differential pressure
line failure. The consequence of releas- and introduces error. Differential pressure can be measured
ing hazardous material from a tubing Calibration of vacuum pressure either with a differential pressure (dP)
failure may warrant the selection of a transmitters is more difficult for in- meter or by subtracting two pressure
Chemical Engineering www.che.com March 2008 49
3FCPJMFS
-5 SFUVSO
'MPXSBUF
HQN
is a temperature delta between the
joined end and the reference junction.
Type J thermocouples, made of iron
and Constantine, are commonly used
#FMPXPGTDBMF
FSSPSJTIJHIGPS
in the CPI for measuring tempera- GMPXTFOTPSBOEEQNFBTVSFNFOU
tures under 1,000°C.
0SJGJDFQMBUFQSFTTVSFESPQ JO)0
RTDs
The second most-common type of tem- Figure 10. Volumetric flowrate is proportional to Figure 11. Due to impulse
the square root of the ∆P, causing high error at less line problems, this "clean" ser-
perature transmitter, RTDs consist of than 10% of span vice did not meet standards
a metal wire or fiber that responds to
a temperature change by changing The rest of the column responded to equipment arrangement for an orifice
its resistance. Though RTDs are less the change in boilup in about 3 min. flowmeter with temperature and pres-
rugged than thermocouples, they are The lag in the control point caused sure compensation.
also more accurate. Typically, they are cycling of the steam flowrate and cre- Typical turndown for orifice plates is
made of platinum. The instrument ated an unstable control loop. The 10:1. Below 10% of span, the measure-
accuracy of thermcouples and RTDs cause was determined to be a thermo- ment is extremely erroneous because
is very good in both. However, ther- couple that was too short for its ther- the volumetric flowrate is proportional
mocouples have a higher error than mowell. Normally, thermocouples are to the square root of the ∆P. At 10% of
RTDs. The total accuracy of a thermo- spring-loaded to ensure that the tip span, the meter is only measuring 1%
couple is 1–2°C. There is greater error is touching the end of the thermow- of the ∆P span (Figure 10).
due to calibration errors and cold-ref- ell, but the instrument mechanics had Multiple meters can be used to
erence junction error. installed a thermocouple of the wrong overcome the turndown ratio when
It is important to note that, with length because they lacked the proper high accuracy is required over the
temperature transmitters, there is replacement part. The poor heat entire span. This is often worth the
a lag in the dynamic response to transfer through the air gap between effort when measuring the flowrate
changes in process temperatures. All the end of the thermocouple and the of raw materials or final products. At
temperature measurements have a thermowell caused the delay in tem- one plant, three orifice plates in par-
slow response, because the mass of perature response. Replacing the in- allel were used to measure the plant-
the thermowell must change in tem- stalled thermocouple with one of the boundary steam flowrate due to the
perature before the thermocouple or proper length fixed the problem. large span and the accuracy required
RTD can see the change. The lag time at the low end of the range. This re-
will depend on the thickness of the Flow sulted in a very complicated system.
thermowell and on the installation. There are many different types of There are many common problems
The thermocouple and RTD must be flowmeters. Here, the types commonly that lead to error in the orifice plate
touching the tip of the thermowell for used in plants will be discussed: orifice measurement, including inaccurate
best performance. If there is an air plates, vortex shedding meters, mag- density, impulse-line problems, erosion
gap between the thermowell and the netic flowmeters and mass flowmeters. of the orifice plate, and an inadequate
measuring device, the heat-transfer number of pipe diameters upstream
resistance of the air will add substan- Orifice plates and downstream of the orifice plate.
tially to the lag time, which is also Orifice plates are the most common An accurate density is required to
why temperature transmitters work type of industrial flowmeter. They are obtain an accurate flowrate. In a plant
better in liquid service. The response inexpensive, but they also have the that has a process feed that varies
time for temperature transmitters greatest error of all the common types from as low as 12% to as high as 30%
in liquid service is between 1–10 s, of flowmeters. Orifice plates measure water, the density changes signifi-
whereas the response time for tem- volumetric flowrate according to the cantly, and therefore an orifice meter
perature transmitters in vapor service following equation: will not provide an accurate reading
is about 30 s. Heat-transfer paste is a 1
¥ $P ´ 2 without density compensation.
thermally conductive silicone grease; Q C v¦ µ Impulse line problems include
it has been used with success in some § R ¶ plugging, freezing due to loss of elec-
plants to improve the response time of Q is the volumetric flowrate, C is a con- tric heat tracing, and leaking. Con-
temperature transmitters. stant, ∆P is the pressure drop across densate filling the impulse lines in
Example. The plant in this example the orifice, and ρ is the fluid density. vapor/gas service and gas bubbles
experienced a temperature lag prob- To obtain an accurate flowrate, an ac- in the impulse lines in liquid service
lem. A thermocouple near the bottom curate fluid density must be known. are also commonly cited. Figure 11
of a large tower controlled the steam Temperature and pressure compen- shows a pipe just upstream of an ori-
to the reboiler. The temperature con- sation are required for vapor or gas fice that was in “clean” water service
trol point had a 10-min delayed re- applications and may be required for two years. There was a filter just
sponse to changes in steam flowrate. for some liquids. Figure 9 shows the upstream of this section of pipe. The
52 Chemical Engineering www.che.com March 2008
impulse lines to the orifice plate flow- the net effect unpredictable. There- for straight runs of pipe upstream and
meter were completely plugged. This fore, under average operating condi- downstream of the meter vary, but
section of pipe was removed and a tions, a typical orifice installation can both are usually longer than for orifice
Teflon-lined magnetic flowmeter was be expected to have an overall inac- plates. In general, 30 pipe diameters
installed instead. curacy in the range of 2 to 5% AR (ac- are required upstream and 15 pipe
Orifice plates can erode, especially tual reading)” [6]. diameters downstream. The upstream
in vapor service with some entrained and downstream piping must be the
liquid. This is common in steam ser- Vortex shedding meters same size pipe as the meter.
vice, and orifice plates should be Vortex shedding meters contain a There are only a few problems com-
checked every three years for wear. bluff body, or a shedder bar, that cre- monly encountered with vortex me-
Orifice plates generally need 20 ates vortices downstream of the object ters. Older models may be sensitive to
pipe diameters upstream and 10 pipe when a fluid flows past it. The meters building vibrations, but newer models
diameters downstream of the orifice utilize the principle that the frequency have overcome this issue. If the shed-
plate for the velocity profile to fully of vortex generation is proportional der bar becomes coated or fouled, the
develop for predictable pressure-drop to the velocity of the fluid. The whis- internal vibration sensor will cease to
measurement. This requirement var- tling sound that wind makes blowing work. This can be avoided by using an
ies with the orifice type and the piping through tree branches demonstrates external vibration sensor. The most
arrangement. This is rarely achieved the same phenomenon. common issue is failing to meet the
in a plant, which introduces error in The fluid’s density and viscosity are Remin requirements over the desired
the measurement. used to set a “k” factor, which is used span. At one plant, every vortex meter
The instrument accuracy of orifice to calculate the fluid velocity from was line-sized, which means it was
plates ranges from ±0.75–2% of the the frequency measurement. The fre- the same size as the surrounding pip-
measured volumetric flowrate. Vari- quency, or vibration, sensor can either ing. The flow went into the laminar
ous problems are encountered with be internal or external to the shedder region in the desired measured range
orifice plate installations, and they bar. The velocity of the fluid is con- in every case. The flow read zero when
have the highest error of all flowme- verted to a mass flowrate using the it transitions to laminar, making the
ters. “Orifice plates are, however, quite fluid density. Therefore, accurate fluid meters useless.
sensitive to a variety of error-induc- density is important for accurate mea- Example. Another good example of
ing conditions. Precision in the bore surements. Vortex meters work well failing to meet the Remin require-
calculations, the quality of the instal- both in liquid and gas service. They ments over the desired span hap-
lation, and the condition of the plate are commonly used in steam service pened on a project where a tower that
itself determine total performance. because they can handle high tem- had been out of service for some time
Installation factors include tap loca- peratures. They are available in many was recommissioned. The distillate
tion and condition, condition of the different materials of construction and flowrate was substantially lower than
process pipe, adequacy of straight can be used in corrosive service. the original tower design and was in
pipe runs, gasket interference, mis- Vortex meters have lower pressure the laminar flow region over the en-
alignment of pipe and orifice bores, drop and higher accuracy than orifice tire operating range. The distillate
and lead line design. Other adverse plates. A minimum Reynolds num- flow was a major control point on the
conditions include the dulling of the ber (Remin) is required to achieve the tower, but the vortex meter could not
sharp edge or nicks caused by corro- manufacturer’s stated accuracy. Vortex read the flowrate. The control strat-
sion or erosion, warpage of the plate meters exhibit non-linear operation egy had to be changed to work around
due to water hammer and dirt, and as they transition from turbulent to this issue until an appropriate meter
grease or secondary phase deposits laminar flow. Typical accuracy above could be installed.
on either orifice surface. Any of the the Remin is 0.65–1.5% of the actual
above conditions can change the ori- reading. In general, the meter size Magnetic flowmeters
fice discharge coefficient by as much must be smaller than the piping size Faraday’s law states that the voltage
as 10%. In combination, these prob- to stay above the Remin throughout induced across any conductor as it
lems can be even more worrisome and the desired span. The requirements moves at right angles through a mag-
Chemical Engineering www.che.com March 2008 53
'5
/PGMPX 'MPX
.BTT
GMPXNFUFS
Figure 15. With an under-
'5
standing of the accuracies of - 18. APRIL
LEIPZIG 14.
mass flowmeters and orifice
flowmeters, we revisit the STAN D G 9
H
question — Which flowmeter IF AT M U N IC
is the most accurate? 5. - 9. M AY
HA LL A2
on your towers. Gather the manufac- to Nick, the following DuPont col- ST AN D 30 3
.PTUMJLFMZUPCFNPTUBDDVSBUF
FWFOBU
MPXGMPXSBUFT*OTFOTJUJWFUPQJQJOHDPO
turer’s information so you can assess leagues contributed their instrument GJHVSBUJPO6OMFTTGMPXJTFSSBUJDEVFUP
the instrument accuracy. Keep in war stories, and the author is grate- GMBTIJOHJOMJOFPSUIFUVCFJTDPSSPEFEPS
mind that the manufacturer’s litera- TFWFSFMZQMVHHFE
UIFPSJGJDFQMBUFTBSF
ful for their willingness to share their
NVDINPSFMJLFMZUPIBWFIJHIFSFSSPS
ture refers to the ideal instrument experiences:
accuracy, which is the accuracy of the • Jim England, DuPont Electronic
measuring device itself. There are Technologies (Circleville, Ohio)
many other factors that contribute • Charles Orrock, DuPont Advanced
to the accuracy of the reading that is Fibers Systems (Richmond, Va.)
displayed on the DCS screen or in the • Adrienne Ashley, DuPont Advanced
data historian. The total accuracy in- Fibers Systems (Richmond, Va.)
cludes the instrument accuracy plus • Joe Flowers, DuPont Engineering
all of the other things that contribute Research & Technology (Wilming-
to error in the measured reading as ton, Del.)
compared to the actual value. Other
inaccuracies lie in digital to analog References
conversions, density errors, piping 1. Gillum, Donald R., Industrial Pressure,
Level and Density Measurement. Resources
configurations, calibration errors, vi- for Measurement and Control Series. ISA,
bration errors, and the list goes on 1995.
and on. Check the field installation 2. Kister, Henry Z., “Distillation Troubleshoot-
ing,” John Wiley & Sons, 2006.
to see what types of problems your 3. Spitzer, David W., Industrial Flow Measure-
meters will experience. ment. Resources for Measurement and Con-
trol Series. ISA, 1990
Get to know your mechanics and in-
4. Trevathan, V. L., editor. A Guide to the Auto-
strumentation experts at your plant. mation Body of Knowledge. ISA, 2006.
Now that you know some of the lingo
of instrumentation, you can better
5. emersonprocess.com/rosemount
6. omega.com
7>D"8:A
converse with your instrument engi- 7. efunda.com hjWbZg\ZYBZbWgVcZBdYjaZ
neers and mechanics. 8. us.endress.com [dgW^dad\^XValVhiZlViZgigZVibZci
9. spiraxsarco.com
Acknowledgements Author
This paper is a compilation of in- Ruth Sands is a senior con-
strumentation basics obtained from sulting engineer for DuPont
Engineering Research &
the references listed below, of trou- Technology (Heat, Mass &
Momentum Transfer Group,
bleshooting experience from many 1007 Market St., B8218,
colleagues at DuPont, and of trou- Wilmington, DE 19898; Phone:
302-774-0016; Fax: 302-774-
bleshooting examples from Henry 2457; Email: ruth.r.sands@
Kister’s most recent book, Distilla- usa.dupont.com). She has
specialized for the last nine MICRODYN
tion Troubleshooting. Much of the years in mass transfer unit TECHNOLOGIES INC
technical information and many of operations: distillation, extraction, absorption,
adsorption, and ion exchange. Her activities P. O. Box 98269
the examples come from Nick Sands, include new designs and retrofits, pilot plant Raleigh, NC. 27624
testing, evaluation of flowsheet alternatives, and
Process Control Leader for DuPont troubleshooting. She has 17 years of experience
Phone 001 - 919 - 341-5936
Chemical Solutions Enterprise in with DuPont, which includes assignments in info@microdyn-nadir.com
process engineering, manufacturing, and corpo-
Deepwater, N.J. Nick has worked for rate recruiting. She holds a B.S.Ch.E. from West
DuPont for 17 years and is a special- Virginia University, is a registered professional WWW.MICRODYN-NADIR.COM
engineer in the state of Delaware, and is a mem-
ist in process control. In addition ber of the FRI Executive Committee.
Circle 31 on p. 76 or go to
adlinks.che.com/7370-31
56x254_biocel_messe_usa_rz.indd 1 07.02.2008 15:36:39 Uhr
T
he precise monitoring and control of The Hall effect is created Limit Switch Options
valve position is essential for efficient au- when a magnetic field is
tomation of both discrete and continuous applied perpendicular to
processes. Measurement of valve position the current direction.
provides the data required for the use of A sensor using the
advanced control strategies and predictive Hall effect is a trans-
maintenance algorithms. ducer that returns a volt-
COM
More effective monitoring of valve posi- age output according to
tion has been an area in which consider- changes in the magnetic
NO
NC
able progress has been made in improving field. For valve position
the performance and reliability of control sensing, an integrated
valves. Modern electrical valve-position Hall-effect sensor and
indicators offer either mechanical or non- magnet assembly detect
contact switching. The position indicators the presence, absence
are typically mounted either directly on a and orientation of a
magnetic trigger. The Mechanical Solid state Reed proximity
valve actuator or work indirectly using a switch sensor switch
non-contact remote feedback device. sensor is powered by
a constant current, and
Mechanical switches develops a varying
Most mechanical-switch valve positioners electrical potential
(Figure, top left) utilize some type of rotary that is proportional to
potentiometer for converting linear to rotary the flux density of a
feedback. These widely used devices are magnetic field applied
similar to variable resistors. perpendicular to the
Rotary potentiometers have an arched coil axis of the sensor.
of wire, over which an arm, called a wiper, Hall-effect proxim-
slides. The wiper is attached to the valve ity sensors used for
cam shaft, and as it moves across the coil of valve positioning offer
wire, a differing voltage output is produced. increased reliability in
The voltage output is proportional to the extreme environments.
angle at which the wiper is oriented. These sensors eliminate
Mechanical switches include contact link- all mechanical contact
ages that are subject to wear over time. The between the valve actua-
wear can eventually degrade performance. tor and the transmitter.
Because there are no
Non-contact proximity positioners moving parts within the
Non-contact technology approaches to Hall-effect sensor and magnet, the life Improved reliability — Safety integrity level
valve positioning can provide accurate expectancy is improved compared to a (SIL) ratings are higher with non-contact
valve-position data without the need for the traditional electromechanical switch. sensors and low-power solenoids. SILs are
linkages or levers required by traditional a measure of safety system performance.
systems. Avoiding mechanical contact in the Reed switches Higher SIL numbers mean better safety
valve positioning system addresses some Some non-contact valve positioners are performance and higher confidence in the
of the performance and cost challenges based on reed switches. A reed switch is field device.
associated with control valves, including an electrical switch that is operated by an Lower costs — Non-contact valve position-
mechanical wear, environmental hazards, applied magnetic field. Reed switches have ers have a lower overall total cost of own-
human error and inaccurate readings. a pair of electrical contacts on ferrous metal ership than conventional devices, thanks
Many non-contact proximity positioners reeds in a hermetically sealed glass enve- to the precise positioning capabilities that
(Figure, top right) incorporate a control- lope. An applied magnetic field moves the can be customized by valve application.
loop feedback mechanism based on an reeds, causing the contacts to either touch Also, the cost of ownership is lowered by
analog PID (proportional integral deriva- or move apart. The contacts can either be ease of calibration and service, and rich
tive) algorithm that has been updated for a open normally, closing when a magnetic diagnostics for predictive maintenance
digital device. The algorithm incorporates field is present, or closed normally, opening signatures.
the Ziegler-Nichols (Z-N) tuning procedure, in the presence of a magnetic field. Bifur- Increased versatility — Non-contact valve
a well-known method for tuning automatic cated reed switches can be used in applica- positioners are designed to be compatible
controllers. It is a two-step tuning approach tions where ultralow power or capacitive with most standard industrial communica-
that adjusts how agressively the valve con- discharge consideration are in effect. tions protocols, including HART, Foundation
troller reacts to errors between the process Fieldbus, AS-I, Modbus, DeviceNet and
variable and the desired setpoint. Benefits Profibus. These devices can help engineers
Significant benefits for non-contact valve take advantage of the cost savings and in-
Hall-effect sensors positioners, include the following: creased diagnostic capabilities of networks,
A number of non-contact proximity position- Greater flexibility — Non-contact position- along with the advantages offered by
ers are based on the solid-state Hall effect, ers utilizing Hall-effect sensors provide improved position sensors.
and are used to help improve monitoring feedback on valve position without link-
and control of production processes. The ages, levers or rotary or linear seals. This Notes
Hall effect refers to a potential differ- allows a remote sensor-head assembly to This edition of “Facts at Your Fingertips” was
ence, known as the Hall voltage, between be mounted a considerable distance from adapted from Jack DiFranco’s article, entitled Ad-
the electronics enclosure, giving engineers vances in Valve Position Monitoring, that appeared
opposite sides of an electrical conductor in the December 2007 issue of Chemical Engineer-
through which an electric current is flowing. increased flexibility and improved safety. ing, pp. 46–50.
Control Valve
Performance
Department Editor: Scott Jenkins
M
inimizing process variability is an response time and over-
important component of a plant’s prof- shoot for a control valve
itability. The performance of control response to a step input
valves within process control loops has a change. In this case,
significant impact on maintaining consistent stem position in percent
processes. This refresher outlines some of the of travel is used as the Output
important aspects of control valve perfor- control valve “output.”
mance, including parameters of both the Step-change size. The
static response and the dynamic response. dynamic response of b
a control valve varies a c
d
Static response depending upon the a < resolution ≤ b
A valve’s static response refers to mea- size of the input step
surements that are made with data points change. Four “ranges” c ≤ dead band < d
Input
recorded when the device is at rest. Key of step sizes to help
Amplitude
static-response parameters for control valves understand the static-
include travel gain, dead band and resolu- and dynamic- response
tion (Figure 1). metrics are defined by
Travel gain (Gx). This term represents the ANSI/ISA standards: Time
change in position of the valve closure • Small input steps (Re- Dynamics are not shown
member divided by the change in input gion 1) that result in no
FIGURE 1. Dead band and resolution, illustrated here, are
signal. Both quantities are expressed as measurable movement
of the closure member key static-response parameters for control valves
a percentage of the full valve span. The
within the specified
closure member is part of the valve trim (the wait time 39
combination of flow-control elements inside Initial overshoot to 38.11 = 23% Final steady-state
• Input step changes average values
a valve). Travel gain measures how well the that are large enough input = 37.84, stem = 37.65
valve system positions its closure member to result in some 38 Stem
compared to the input signal it receives. control-valve response
Input, stem %
S
ince its inception in the 1960s
and its first official publica-
tion in 1977, the Hazard and
Operability Study (HAZOP)
has become one of the most power-
ful tools for identifying process haz-
ards in the chemical process indus-
tries (CPI). Utilizing systems that are
qualitative or even simplified semi-
quantitative, the HAZOP method has
been increasingly used, not only as a
tool for identifying process hazards,
equipment deficiencies or failures and
operability problems and assessing
their risks, but also as a tool for priori-
tizing actions and recommendations
for process-risk reduction. Reducing
risk is especially important in ensur-
ing the safety of the personnel who
must work in the plant environment
each day (Figure 1).
The HAZOP methodology is a sys-
tematic team-based technique that
can be used to effectively identify
and analyze the risks of potentially
hazardous process operations. It is
the most widely used process hazard
analysis (PHA) technique in numer-
ous industries worldwide, including
petrochemicals, pharmaceuticals,
oil-and-gas and nuclear, and is used
during the design stages of new pro-
cesses or projects, for major process
modifications and for periodic review
of existing operations. FIGURE 1. HAZOP studies are useful tools in reducing process risk, and they provide safeguards against
A HAZOP is a time-consuming ex- hazardous scenarios for the personnel who must maintain and operate the plant
ercise and should be conducted in
such a way to ensure that the results to correct these mistakes if or when Benefits of a HAZOP
justify the effort. This article presents they occur during the course of the The advantages offered by HAZOP
some common mistakes that can HAZOP study. Therefore, the selec- over other process-risk analysis
jeopardize a HAZOP team’s task. Fre- tion of an experienced facilitator is tools are numerous, and include
quent or chronic occurrence of these an essential element for assuring the the following:
mistakes indicates potential gaps in success of the HAZOP. Without an • It is a rigorous process; it is structured,
the site’s process-management sys- adequate depth of knowledge and systematic and comprehensive
tem. However, it is ultimately the re- experience, the HAZOP can become • It is adaptable to the majority of CPI
sponsibility of the HAZOP facilitator a “check the box” exercise. and manufacturing operations, in-
54 Chemical Engineering www.chemengonline.com december 2015
cluding those in petroleum refiner-
ies (Figure 2) and other oil-and-gas
processing plants, nuclear facili-
ties, and specialty chemical, phar-
maceutical and even high-speed
manufacturing plants
• It is team-based and allows the
interchange of knowledge and ex-
perience between the participants
• It helps to anticipate potential ac-
cidents or harm to employees, the
facility, the environment and the
surrounding community
• It functions as a type of training for
the team’s participants and leader,
who are required by the nature of
the method to look at the process
from a new perspective — not just
from the perspective of “how should
it run?,” but also “how can it fail to FIGURE 2. Many processes in the CPI are potentially hazardous if not managed correctly. HAZOP studies
run correctly?” seek to prioritize actions to reduce process risks, and are adaptable across a wide range of
A HAZOP is time-consuming be- industrial sectors
cause it requires the participation estimate of the time required for the Mistake 3: Incorrect size of HAZOP
of a multi-disciplinary team over ex- team based on the process descrip- team. The HAZOP team should be
tended timeframes. This investment tion and preliminary count of HAZOP limited in size, ideally five to seven
of time and personnel, often involv- nodes (specific portions or topics of people, excluding the HAZOP facili-
ing third parties, means that the per- the study process) so that managers tator and the HAZOP scribe or sec-
formance of the HAZOP needs to be are aware of the degree of personnel retary. A team that is too large can
optimized to maximize its value. The commitment that will be required. easily lose focus, dwell on a subject
following sections detail some com- Mistake 2: Incomplete, inaccurate or issue too long, or be disruptive. It
monly found mistakes that occur or unavailable process safety in- is human nature that all participants
during the planning, execution and formation. Another common mis- seek to present their perspectives,
followup stages of a HAZOP. take during a HAZOP is not having but this can lead to excessive dis-
all the prerequisite process safety cussion. A group that is too small will
Planning stage information (PSI) and other valuable not likely include the right expertise
Mistake 1: Mismanagement of information available, including out- or provide enough different perspec-
time-allotment issues. One of the of-date or incomplete information. tives to evaluate the process haz-
most frequent mistakes of a HAZOP This is especially critical regarding ards and controls adequately or in
is failure to manage the time allot- piping and instrumentation diagrams the right detail.
ted for the study. A HAZOP is often (P&IDs), current standard operating
scheduled for a set amount of time, procedures (SOPs) and appropriate Execution stage
neither by the HAZOP facilitator nor data on flammability, combustibility, Mistake 4: Lack of focus during
the team, and sufficient time may not reactivity, toxicity and electrostatic the meeting. A HAZOP is a com-
have been allocated. Furthermore, properties of materials in all forms plex exercise that requires the con-
there may be little or no flexibility in and phases, as well as compatibil- centrated and coordinated contribu-
the schedule. An insufficient amount ity of chemicals with each other and tion of all the members of the team.
of time for the HAZOP limits discus- with the processing equipment. If the Distractions should be minimized
sion and brainstorming and reduces HAZOP is conducted by an external in order to ensure and maintain the
the quality of the analysis, in turn facilitator, it is the responsibility of the team’s focus. Therefore, team mem-
leading to some of the mistakes dis- owner of the process to verify the in- bers should not be allowed to come
cussed in more detail below. tegrity of the PSI. and go into and out of the meeting,
Estimating the duration of a Related to this, it is not acceptable take phone calls, answer emails,
HAZOP is not an exact science, and that participants attend the HAZOP or discuss issues not related to the
it requires a good knowledge of the for the purpose of obtaining informa- HAZOP during the sessions. Use of
methodology, the complexity of the tion on a process or project. HAZOP an offsite venue may be helpful to
process, the nature of the risks that participants should be well prepared prevent plant operations from be-
can be identified up front and the id- to contribute to the discussion and coming a distraction.
iosyncrasies of the group. Although a have all requisite background infor- It is the responsibility of the HAZOP
HAZOP should not be open-ended in mation with them. It is the respon- facilitator to maintain the focus of the
time allotment, the ideal HAZOP has sibility of the facilitator to instruct all group and keep the HAZOP process
some flexibility built into the sched- participants that they must come to moving by allowing some open dis-
ule. The team leader should make an the HAZOP prepared. cussion on the issue, node and con-
T
500
he chemical process industries (CPI) are func- 530 mm (min)
60
tioning in an era of globalization, and between the
400
prevailing economic conditions and upheavals 50
in the energy sector, the number of new invest-
Efficiency, %
Head, m
ments in CPI facilities has fallen in recent years. Many 300 40
industries are seeking cost reductions by revamping
existing plants with minimum investment. The objective 30
is to reduce the cost of production through the use of 200
upgrades and new technologies, to remain competitive 20
in the market. By way of example, if one wants to set 100
up a new complex to produce ammonia and urea, the NPSHR 10
specific capital cost will be on the order of $666/ton of
urea. By comparison, if an existing plant is revamped to 0 0
0 250 500 750 1,000 1,250 1,500 1,750
raise the existing production from 100% to 120% (that Capacity, m3
is, adding 20% additional capacity), this can be done at
an expenditure that is closer to $300/ton to achieve this FIGURE 1. Shown here are typical pump characteristic curves, with three
different impeller sizes, showing capacity versus head, and NPSHR versus
incremental production capacity
This article reviews key concepts, objectives and pro-
cedures that are needed to successfully carry out vari- a turnkey basis. Meticulous planning related to the hook-
ous types of CPI plant revamps. up of tie-in points arising out of expansion schemes can
help to reduce the amount of downtime required to ex-
The need for revamps ecute the revamp schemes and put the plant back on-
Chemical process plant revamps are typically under- line.
taken for the following reasons:
• To change in feedstock composition Targeted revamp capacity, change in process
• To adopt energy-conserving processes in light of in- In general, it is possible to increase the rated capacity
creasing energy costs of a plant by 10%, with very little added expenditure.
• To reduce the fixed-cost components of production, But to increase capacity by 20–50% over the nameplate
by increasing capacity within the existing facility capacity, substantial modifications must be taken into
• To extend the life of a well-maintained process plant consideration that often involve implementing different
Similarly, there are many benefits to conducting an ap- technologies from the ones already applied in the exist-
propriate plant revamp. These include the ability to: ing plant. When seeking such notable increases in pro-
• Increase the reliability of equipment, leading to reduced duction capacity, plant operators and managers must
downtime and maintenance costs not only verify the soundness of the economics, but also
• Reduce energy consumption carefully evaluate the potential drawbacks, if any.
• Extend useful plant life Sometimes the existing process path may have to be
• Reduce the cost of production, thereby improving the changed to enhance the capacity of the plant, since the
overall bottom line for the facility current process may not yield the desired efficiency or
However, experience shows that inefficient implementa- conversion rates. Two cases are discussed below.
tion of proposed revamp options can lead to failure, so Example 1. In the case of units to recover liquefied pe-
care must be taken to avoid this by building the right troleum gas (LPG) from natural gas, such units are de-
team of experts. This team typically includes represen- signed for a certain composition of feed gas. The need
tatives of the process licensor company, engineering for a revamp often arises if the gas composition has
and project-management consultants, and experts from changed and the expected recovery of C3/C4 and higher
the owner company representing diverse fields, such as compounds has become unprofitable. In this case, the
operations, project management and maintenance. If expected recovery of LPG and natural gas liquids (NGLs)
sufficient expertise for the proposed revamp is not avail- can be achieved by compressing the feedstock to higher
able internally, one can hire consultants to carry out the pressures than present levels, or by spiking heavier NGLs
feasibility studies and implementation of the revamp on back to the feed gas stream. Thus, such a revamp re-
48 Chemical Engineering www.chemengonline.com december 2015
quires a study to assess the technical and economic fea- raw materials, utilities and energy per unit of production
sibility of the different process paths being considered. are tabulated. The material-and-energy balance of the
Example 2. A feedstock change from naphtha to natural existing operation, and the required revamp plant load,
gas in ammonia plants, hydrogen plants and methanol are prepared.
plants also necessitates a need for revamp of the re- The existing equipment components are rated for
former section and front end, but in many cases, the the revamp conditions, and then changes and required
existing process path can be retained. In this case, the new equipment are identified. Cost estimates of various
absorbed duty of the reformer — which tends to be the schemes are prepared (after consultation with various
major energy-consuming equipment found in the system vendors). Feasibility studies, followed by detailed project
— and the burner duties required vis-a-vis the required reports (DPR), are also prepared. The potential rates of
reformer absorbed duty are calculated to check their suit- return of various options are studied. The best option
ability. The maximum skin temperature of the reformer available (on the grounds of economic sustainability and
tubes for the feedstock change must be checked. technical feasibility) is then selected, so that the basic
In all cases, the existing process path, along with other engineering design package (BEDP) can be prepared,
options, must be studied in detail to arrive at the most and the revamp project implemented.
economical and technically feasible revamp option. As noted, successful revamps require assembling
the right revamp team. Typically, such a team consists
Lifecycle of the plant of individuals from the process licensor company, con-
The different phases of a plant’s lifecycle must be taken sultants for basic engineering and detailed engineering
into consideration when planning a revamp. Such phases services, contractors for specific electrical-, mechanical-
include the following: and instrumentation-related aspects of the project, and
1. Incubation stage — Initial stabilization period various engineers from the owner’s group (for instance,
2. Growth stage — Optimization and debottlenecking of those who represent specific disciplines and have a con-
operations to improve the efficiency crete understanding of the current operation).
3. Maturity stage — Attainment of stable operation The following planning steps should be undertaken:
4. Declining stage — Realization that plant capacity is 1. Estimate the plant’s inherent capacity from past and re-
not sustainable because of frequent equipment fail- cent data. This can be done by identifying weak areas
ures or excessive maintenance requirements in the plant (for instance, those that are contributing
Revamping the plant during Phases 1, 2 or 3 is relatively to non-realization of rated or required plant capacity),
easy, whereas revamping a plant during Phase 4, when or by conducting an end-to-end survey of the plant.
the facility is already in decline, requires the engineering Once such a study is carried out, efforts should be
team to adapt many of the modern technology options made to predict the potential performance improve-
to an aging infrastructure, and to replace many equip- ments of the plant if the weak areas are rectified.
ment components. 2. Prepare the process scheme and the equipment data
sheets. Carry out feasibility studies of all options (in-
Objectives of a revamp cluding both technical and financial aspects of the pro-
The objectives of a plant revamp should be spelled out posed revamps) and then develop the detailed project
prior to studying the options. Possible objectives could report. Set the target of the revamp in terms of time
be the following: (schedule) and cost.
• Enhance capacity from the present operating level to 3. Implement the approved revamp. Ideally, the revamp
expand capacity to, say, 110%, 120%, 130% of rated activities should be carried out during the annual
capacity scheduled turnaround period for the plant, to minimize
• Reduce production costs unscheduled downtime.
• Reduce pollution
• Reduce the consumption ratios of various raw materi- Estimate plant capacity
als and utilities Many older CPI plants can run at or above the rated ca-
• Reduce maintenance costs and increase the onstream pacity continuously for a week or a month. But due to cer-
factor tain operating limitations, and downtime that may arise
• Upgrade the technology to keep pace with the new from some underperforming equipment, the annual rated
developments, and to increase the plant life capacity is seldom achieved. Analyzing past operating
• Minimize plant shutdown data on a monthly basis (for the past 10 years or so) will
These objectives can be achieved by maximizing efficiency, reveal which equipment components are most often to
yield and conversion of raw materials in various sections. blame for downtime, and are thus affecting overall capac-
Specifically, plant revamps are often implemented to im- ity utilization. Such a study of past data is often called a
prove process optimization, increase energy conserva- weak-area analysis. Similarly, sometimes an end-to-end
tion, improve product quality and expand capacity. survey of the plant (from the plant commissioning to the
present day) is also conducted.
Key revamp procedures Existing equipment poses both opportunities (in the
Every revamp project should start by identifying the goals form of underutilized capabilities) and challenges (in terms
and actual bottlenecks. A material-and-energy balance of limitations). The ability to identify problem areas can
for the base case should be developed to reflect the ac- help the team to prioritize their debottlenecking efforts in
tual operating conditions. The consumption of various order to improve capacity utilization more quickly.
In Brief
The need for point
level switches
Types of point-level
switches
Summary
S
afety is an important and common Figure 1. Preventing overfilling of chemical storage tanks
subject of discussion in the chemi- requires proper selection of high-high point-level switches
cal process industries (CPI) today. model. Determining what may go wrong and
Conversations on safety include then taking steps to reduce the possibility by
many topics, such as risk assessment, risk adding safety procedures, retention dikes,
mitigation, and tolerable risk. Acronyms safety instrumentation and so on, are univer-
like SIS (safety instrumented systems), SIL sal to any safety program.
(safety integrity level) and PFD (probability of The following are assessments of point
failure on demand) and others have become level switches as they are used in over-
part of the safety lexicon in CPI facilities fill prevention safety programs. We review
throughout North America and the world. All some basic concepts and look at some of
of these terms and acronyms can be con- the common technologies used to prevent
fusing, complicating what steps need to be overfilling of vessels (Figure 1). The positive
taken to make a facility safe. and negative aspects of each technology are
Regardless of how the safety concepts are also considered.
labeled, there are a few principles that form
the basis for all safety models. Whether you The need for point-level switches
subscribe completely to the SIS concept or Point-level switches are often used in appli-
have developed your own safety procedures cations designed to prevent accidents. In-
internally, risk assessment and risk mitiga- dustries that manufacture or store materials
tion are the two key concepts in any safety that are potentially hazardous employ point-
ROTOFORM
GRANULATION
FOR PETROCHEMICALS
Figure 3. The ability to test in-situ validates the functionality of the safety switch while reducing main-
tenance and downtime
AND OLEOCHEMICALS
switches, as the name implies, utilize parts and the potential for a lack of
a float that changes position due to buoyancy are critical failure points.
buoyancy and indicates presence From the standpoint of safety appli-
of a liquid. The float may move on a cations, floats should be avoided.
vertical shaft and trip a magnetically Ultrasonic gap point-level switch.
coupled reed switch or may pivot Ultrasonic gap switches are com-
on an access providing a mechani- prised of two piezoelectric crystals
cal internal switch to activate. The situated on opposite sides of a gap.
appeal of float switches is that they One crystal is excited electrically and High productivity solidification of
are simple devices and relatively in- generates acoustic energy that is products as different as resins, hot
expensive. However, the mechani- directed across the gap toward the melts, waxes, fat chemicals and
cal nature of a float and the moving second crystal. With air or gas in the caprolactam has made Rotoform® the
granulation system of choice for
parts that can hang up or bind due to gap, the energy is not strong enough
chemical processors the world over.
coatings makes them questionable to reach the second crystal. Once Whatever your solidification
for use in safety applications. the gap fills with a liquid the acoustic requirements, choose Rotoform for
The ability to test a float switch is energy is coupled through the liquid reliable, proven performance and a
also suspect. Some manufacturers molecules, reaches the second crys- premium quality end product.
provide a lift arm to physically move tal and completes the circuit, indicat-
the float to make it change state ing that the liquid is present. High productivity –
from normal to alarm. This test is Ultrasonic gap switches have no on-stream factor of 96%
insufficient to exercise potential fail- moving parts to wear or hang up, Proven Rotoform technology –
ures, such as leaking floats, and may which is an advantage over mechan- nearly 2000 systems installed
in 30+ years
not identify binding or heavy coat- ical switches, such as floats. How-
Complete process lines or
ings. Some test arms are fitted with ever, materials that leave coatings retrofit of existing equipment
magnets that will release if the float and materials that have suspended Global service / spare parts supply
is heavy due to leakage or coating, solids, or are aerated, will block the
but even this precaution is suspect. acoustic energy, causing a failure.
As such, the only true way to test In-situ testing of ultrasonic gap
the float switch is to remove it from switches that validate all potential
the vessel for testing, or to raise the failures is not possible. Some manu-
product level to the high-high switch, facturers provide test buttons that Sandvik Process Systems
which, as previously discussed, is are used to test the switch. This Division of Sandvik Materials Technology Deutschland GmbH
Salierstr. 35, 70736 Fellbach, Germany
not permitted. test operates in one of two ways. In Tel: +49 711 5105-0 · Fax: +49 711 5105-152
info.spsde@sandvik.com
Floats are best suited for simple some products, there is a second set www.processsystems.sandvik.com
non-critical applications. Moving of crystals that are wired together.
Circle 18 on p. 190 or go to adlinks.chemengonline.com/56204-18
Alberto Ferrari
Figure 1. Integrated system architecture is illustrated here for
Ferrari Granulati an automated mill with an online particle-size analyzer
T
he pursuit of manufacturing This article draws on experience from turnkey integration within a so-
excellence is a strong theme from different industries as they apply phisticated control platform to sensor-
across all of the chemical pro- realtime particle-sizing technology, to only purchase for the inhouse imple-
cess industries (CPI). Improv- demonstrate the various control strat- mentation of simple closed-loop control.
ing operations by eliminating vari- egies that it supports and the benefits This ready availability of realtime mea-
ability and waste, a cornerstone of the that can be derived. surement presents an important oppor-
six sigma approach, is essential for Note: With an online system the tunity to realize a number of important
competitive performance in the global analyzer is typically installed on a practical gains, even before considering
marketplace. In addition, of course, dedicated loop fed from the process the issue of process control.
health, safety and the environment line, while with an inline instrument Full automation — from sample
(HSE) remains the subject of intense the analyzer sits directly in the bulk extraction through the delivery of
scrutiny and concern. Within this cli- process flow. Both approaches provide results to a control system — means
mate, automation of both analysis and continuous realtime data so for the that online analysis completely elimi-
control, can be highly advantageous. purposes of this article the term on- nates the issue of operator-to-operator
Particle size analysis is a measure- line has been used to cover both types variability. In addition, the ability to
ment technique that has already suc- of installation. analyze a much higher proportion of
cessfully completed the transition from the process stream improves the sta-
laboratory to line and become an estab- Moving toward realtime tistical relevance of the data. Equally
lished part of the automation toolkit. It should be stated from the outset important is that the benefits of au-
Reliable, online particle-size analysis that offline analytical capability re- tomation come with a dramatic de-
systems are now commercially proven mains vital. Essential during research, crease in the amount of manual labor
across a number of sectors, on both it is also frequently the norm for final required for analysis, and a significant
wet and dry process streams, with all quality control checks. Furthermore, reduction in the containment and ex-
aspects of project implementation well- certain techniques that yield critical posure risks associated with manual
scoped and understood. With this mat- information have yet to move success- sampling and analysis.
uration has come greater accessibility fully to the process environment. That Investing in realtime measurement
to sectors that previously faced tech- said, for routine process monitoring, technology may be justifiable simply
nical or financial barriers to adoption. an offline regime is far from ideal and on the basis of these practical ben-
Requirements are different in each online measurement, if available may efits, especially where the envisioned
and every case, from a simple single be the preferred option. set-up is relatively simple. The cost
PID (proportional integral derivative) Laser-diffraction particle sizing ex- of installation is offset by savings in
control loop to multivariate statistical emplifies a number of commercially manpower, and the operational team
control or the rigors of operation within proven, online analytical technologies maintains manual control exactly as
a highly regulated environment. Nev- with established credentials for real- before, using the online system solely
ertheless, the building blocks needed to time monitoring. Today it is possible to for particle size monitoring. While this
fashion an optimal solution are there, tailor an online particle-sizing solution approach may deliver improved opera-
for the vast majority of applications. that closely matches user requirements, tions and variable cost gains, it fails
Chemical Engineering www.che.com February 2012 41
Solids Processing
Figure 2. For
Average size and transmission
simple, closed-
2,600 3,600 loop control,
2,800 rpm rpm
rpm 3,200 online analysis
rpm 3,400
rpm 3,800 can be used to
4,000 rpm track particle
rpm
150.00 250.00 size changes that
are induced by
varying mill rotor
Particle diameter, µm
112.50 187.50 speed over time
Transmission, %
Avg{Dv(90)} = 170.54
75.00 125.00
Avg{Dv(50)} = 57.58
37.50 62.50
Avg{Dv(10)} = 9.16
0.00 0.00
11/13/2008 -17:05:00 11/13/2008 -17:15:00 11/13/2008 -17:25:00 11/13/2008 -17:35:00 11/13/2008 -17:45:00
to fully realize the potential that on- Case study: Simple closed loop con- manipulation (Figure 2). A 30-s rolling
line instrumentation offers in opening trol on the basis of realtime data. average Dv50 (median particle size)
up the route to automated control. An An automated mill system recently in- figure from the online particle-size
efficient control-automation strategy stalled at a commercial pharmaceutical analyzer drives this loop.
fully exploits the information stream manufacturing site is shown in Figure To test the response of the system,
provided by realtime measurement 1. It was developed as a widely appli- the setpoint for the loop was reduced
and maximizes return on investment. cable, validated alternative to manual from an initial 58 microns to 50 mi-
mill control using offline particle-size crons, and then back up to the original
Simple closed-loop control measurement. One routine operation value. Despite the absence of compre-
The simplest option when implement- in the company is milling an active hensive loop tuning — proportional
ing automated control on the basis of pharmaceutical ingredient (API), typi- (P) only control was used at this stage
continuous data from an online in- cally recovered through crystallization, — the results were good. Steady op-
strument is usually a single variable to a defined particle size. The particle eration at 50 microns was established
PID control loop. Such an approach size distribution of an API is often a just 30 s after the change was made,
can prove highly productive and be critical quality attribute because of its and the final transition was complete
an efficient way of automating exist- impact on clinical efficacy and drug in under 2 min.
ing manual control strategies. Even product manufacturability. To mill a new batch with this sys-
the implementation of one automated The comminution mill has fast dy- tem, the operator simply selects the
loop changes the process from being namics, making rapid and continuous target particle size and feeds mate-
fixed to becoming responsive. A fixed data acquisition and interpretation es- rial into the mill. Control is then suf-
process translates, or magnifies, up- sential. Since the selected online par- ficiently tight to largely eliminate
stream variations on to the product or ticle-size analyzer has a measurement the production of out-of-specification
downstream process; a responsive one rate of four complete particle-size dis- material. Contrasting this control
either erases or reduces their impact. tributions per second, it can efficiently scheme with offline or manual analy-
For milling, a common approach is track even this swiftly changing pro- sis quickly brings the multiple ben-
to automatically vary either mill speed cess in fine detail. efits into sharp focus:
or downstream separator variables This fully integrated system uses up- • When using offline particle-size
(classifier speed, for example) to meet graded programmable-logic-controller analysis, each new batch required a
the product specification (as in the ex- (PLC) code and proprietary software potentially lengthy iterative process
ample below). An exactly analogous to handle data exchange between the to determine the appropriate rotor
strategy in emulsification processes main hardware units. The operator speed, in order to meet the defined
allows key variables, such as pres- interacts with the central controlling specification. Eliminating this step
sure, to be automatically manipulated PC via the mill HMI (human machine has saved both time and material.
in order to control droplet size. Since interface) and can do the following: • Prior to automation, the rotor
online particle-size analyzers are com- input set points; start and stop the speed was fixed for each batch,
mercially available for both wet and mill or analyzer remotely; perform based on a test sample. However, if
dry process streams, any unit involv- background tests; and receive particle the sample was not representative
ing comminution to a defined particle size results. A closed control loop links of the batch, or if segregation had
size can potentially benefit from this particle size with mill rotor speed, the occurred, the rotor speed would be
very basic type of automation. principal operating parameter for size less than optimal and the product
42 Chemical Engineering www.che.com February 2012
Particle High
size efficiency
product separator
quality
Separator the future this should ease the inte-
speed
To gration of both process and laboratory
storage
silo
systems. Enabling software, based on
this specification, is already avail-
Elevator From able commercially. These advances
power Exit water spray clinker
air flowrate silos will reduce the difficulty and cost of
(separator
feedrate) Elevator implementing customized multivari-
conveyor Feed ate control strategies, bringing the
rate potential rewards within the reach of
Exit
temperature more manufacturers.
Ball mill Case study: Multivariate control
LEGEND
of a heavy commodity milling cir-
Controlled variables
Manipulated variables
cuit. In 2006, Vulcan Materials Com-
pany (Birmingham, Ala.) made the
Florida Rock Industries – Thompson Baker Cement Plant (Vulcan Materials Co.) decision to transform control of its ce-
ment finishing circuit (Figure 3). The
Figure 3. This upgrade to a cement finishing circuit for a milling clinker (and gypsum project involved the three following
and limestone) helps achieve the fineness required to meet final cement specifications significant changes:
• Switching from Blaine measurement
inconsistent. Since the automated ensure that the total feedrate to the to laser-diffraction particle-size anal-
mill responds to, and compensates mill remains constant. In combina- ysis, with the intention of more pre-
for, any variability in the feed tion with the fixed operating strategy cisely targeting cement performance
stock, the product particle size has these loops ensure that exceptional and accessing online technology
become extremely consistent. product quality is achieved at com- • Adopting online, rather than offline
petitive cost. analysis for process control
Multiple control loops As with the previous example, the • Selecting and installing a powerful
The next step beyond a single control architecture of the control loops em- model-predictive-control package to
loop involves multiple, simple control ployed here is relatively simple: one automate process control
loops, which enable the parallel ma- single process variable manipulated Vulcan Materials installed a propri-
nipulation of a number of variables to on the basis of a single measured etary solution for multivariate con-
simultaneously meet product quality, variable. Nevertheless, they are trol and an online laser-diffraction
variable cost and throughput goals. For highly effective with each loop effi- particle-size analyzer. At the heart
example, at its plant close to Verona in ciently targeting one specific aspect of the control package is a multivari-
Italy, Ferrari Granulati mills very fine of process performance. ate process model that is tuned using
marble powders of exemplary quality plant data to accurately predict plant
[1]. Three discrete products are mar- Multivariate process control performance from a range of inputs.
keted with Dv50s in the range of three Just as a seasoned operator bases Automatic manipulation of process
to eight microns. Here, online particle- manual control decisions on every variables, on the basis of these predic-
size analysis has been used extensively piece of relevant information, the tions, achieves plant performance tar-
to develop the design of the mill and a most sophisticated automated control gets, which are as follows:
control strategy for the milling circuit relies on an array of data, rather than • Maintain product quality
(mill and associated classifier). a single input. Multivariate process • Reduce variability and improve op-
The adopted strategy relies largely control systems take in and use data erational stability
on fixing a number of process variables, from a number of sources and, in com- • Maximize fresh feedrate subject to
at values defined through detailed bination with a process model, provide equipment constraints
optimization trials that are based on multiple outputs, simultaneously ma- The process model runs in real time,
realtime measurement. These values nipulating various operational param- employing an integrated steady-state
have been defined for each product. eters. Such systems work within well- optimizer and dynamic controller to
Two independent, automated control defined boundaries to target optimal drive the system toward optimal op-
loops are, however, applied to opti- operation at all times. eration within the above constraints.
mize mill performance and maximize Quite recently, steps have been The impact of changing manipulated
plant throughput on an ongoing basis. taken to reduce one of the barriers to variables is projected into the future;
One loop maintains a prescribed pow- implementing multivariate process predictive control ensures that mul-
der depth on the table of the vertical control: the difficulty of integrating tiple performance targets are met
roller mill to ensure efficient commi- analyzers from different suppliers. simultaneously and that process out-
nution and prevent excessive wear of The new OPC Foundation Analyzer puts are as close as possible to desired
the mill. The other controls the rate of Device Integration (ADI) specifica- reference trajectories. Optimization
fresh feed to the mill with reference to tion [2] provides a common standard procedures are repeated each time
the recycle rate from the classifier, to for instrument manufacturers. Into process values are re-read, following
Chemical Engineering www.che.com February 2012 43
Solids Processing
a change, in order to maintain the fu- The installed solution has pushed operation presents an opportunity for
ture prediction-horizon period. This the circuit into a new operating re- improved control, reduced risk and fi-
is termed “realtime receding horizon gime, uncovering better control strate- nancial gain. For online particle-size
control” and is characteristic of model gies than had been identified through analysis proven automated systems
predictive control. manual operation. Significant benefits have developed to the point of wide-
Here, a number of loops are operat- have accrued, including the following: spread availability, making this op-
ing in tandem (Figure 3). Manipulat- • A 20% reduction in specific energy portunity financially and technically
ing clinker feedrate and separator consumption feasible across a broad range of manu-
speed controls product quality. These • An improvement of 15% in one-day facturing sectors. Options for invest-
same two parameters are also used, strength levels (a primary perfor- ment now range from fully integrated
together with air flow through the mance indicator) turnkey solutions to sensor-only pur-
mill, to drive separator feedrate (mea- • An increase in throughput in excess chase for in-house implementation.
sured as elevator power) toward a de- of 15% Continuous, realtime analysis of a
fined high limit. Air flowrate through These savings are not individually at- critical variable, one that directly influ-
the mill and exit water spray vari- tributable to either the analyzer or the ences product performance, provides a
ables control the temperature of exit- control package but arise from symbi- platform for developing and imple-
ing material. Finally, a stabilizing loop osis between the two, which has un- menting the very best automated con-
minimizes a function defined as “mill locked the full potential of each. The trol strategies for a given application.
condition”, calculated from the rate of payback time for the entire project is Such strategies deliver multiple eco-
change of elevator and mill power. The estimated at just over one year, based nomic benefits in the form of reduced
online particle-size analyzer measures on energy savings alone. waste, increased throughput, enhanced
product quality in real time, providing product quality, reduced manual input
vital information that is used by the Conclusions and lower energy consumption. Auto-
model in combination with an array of The automation of process analysis mating control realizes the full poten-
other process measurements. and its closer integration with plant tial of realtime analysis, extracting
maximum return from an investment
in process analytical technology. ■
Edited by Rebekkah Marshall
D
ifferent methodologies are Table: Different PHA methods and Approaches
available for conducting the Method Description
structured reviews known as
Consequence analysis This method quantitatively assesses the consequences of hazardous material
process hazards analyses (PHAs) for releases. Release rates are calculated for the worst case and also for alterna-
new processes. PHAs are often con- tive scenarios. Toxicological endpoints are defined, and possible release dura-
ducted or moderated by specialists, tion is determined
with participation by the design team,
representatives of the facility owner, Hazard identification HAZID is a preliminary study that is performed in early project stages when
and experienced process operators. analysis (HAZID) potentially hazardous materials, general process information, initial flow diagram
and plant location are known. HAZID is also generally used later on to perform
Each different PHA method is bet- other hazard studies and to design the preliminary piping and instrumentation
ter-suited to a specific purpose and diagrams (P&IDs)
should be applied at different stag-
es of the project development. The What-if method The what-if method is a brainstorming technique that uses questions starting
table includes brief descriptions of with “What if...,” such as “What if the pump stops running” or “What if the op-
some of the most widely used PHA erator opens or closes a certain valve?” For best results, these analyses should
be held by experienced staff to be able to foresee possible failures and identify
methods in the chemical process in- design alternatives to avoid them
dustries (CPI).
Hazard and operability The HAZOP technique has been a standard since the 1960s in the chemical,
When to use different methods study (HAZOP) petroleum refining and oil-and-gas industries. It is based on the assumption
Different types of PHA studies have that there will be no hazard if the plant is operated within the design param-
varying impact, depending on the de- eters, and analyzes deviations of the design variables that might lead to unde-
sirable consequences for people, equipment, environment, plant operations or
sign phase in which they are applied. company image.
For example, if a consequence analy- If a deviation is plausible, its consequences and probability of occurrence
sis is not performed in a conceptual are then studied by the HAZOP team. Usually an external company is hired to
interact with the operator company and the engineering company to perform
or pre-FEED (front-end engineering this study. There are at least two methods using matrices to evaluate the risk
and design) phase, important plot- (R): one evaluates consequence level (C) times frequency (F) of occurrence;
plan considerations can be missed, and the other incorporates exposition (E) as a time value and probability (P)
such as the need to own more land ranging from practically impossible to almost sure to happen. In this method,
the risk is found by the following equation: R = E × P × C
to avoid effects on public spaces; or
the fact that the location might have a
Layer-of-protection The LOPA method analyzes the probability of failure of independent protection
different elevation with respect to sea analysis (LOPA). layers (IPLs) in the event of a scenario previously studied in a quantitative hazard
level than surrounding public places evaluation like a HAZOP. LOPA is used when a plant uses instrumentation inde-
impacted by a flare plume. pendent from operation, safety instrumented systems (SIS) to assure a certain
Some other studies, like HAZOP, safety integrity level (SIL). The study uses a fault tree to study the probability of
failure on demand (PFD) and assigns a required SIL to a specific instrumentation
cannot be developed without a con- node. For example, in petroleum refineries, most companies will maintain a SIL
trol philosophy or piping and instru- equal to or less than 2 (average probability of failure on demand ≥10−3 to <10−2),
mentation diagrams (P&IDs), and are and a nuclear plant will tolerate a SIL 4 (average probability of failure on demand
≥10−5 to <10−4)
performed at the end of the FEED
stage or at the end of the detailed
engineering phase (or for improved Fault-tree analysis Fault-tree analysis is a deductive technique that uses Boolean logic symbols (that
is, AND or OR gates) to break down the causes of a top event into basic equip-
results, at the end of both) to define ment failures or human errors. The immediate causes of the top event are called
and validate the location of pressure “fault causes.” The resulting fault-tree model displays the logical relationship
safety valves (PSVs) as well as to between the basic events and the selected top event
validate other process controls and
instrument safety requirements. Quantitative risk assess- QRA is the systematic development of numerical estimates of the expected
QRA or LOPA evaluations (or both) ment (QRA) frequency and consequence of potential accidents based on engineering
evaluation and mathematical techniques. The numerical estimates can vary
are undertaken after the HAZOP study from simple values of probability or frequency of an event occurring based on
to validate siting and define safety in- relevant historical data of the industry or other available data, to very detailed
tegrity levels (SIL), to finally meet the frequency modeling techniques. The events studied are the release of a haz-
ardous or toxic material, explosions or boiling-liquid expanded-vapor explosion
level required by the plant. n (BLEVE). The results of this study are usually shown on top of the plot plan
Level 1
Sebastiano Giardinella causes
Inelectra S.A.C.A.
Improperly
R
Lack of a relief Undersized relief installed relief Miscellaneous
elief systems are the last line of device device device deficiencies
defense for chemical process fa-
cilities. Verifying their capabil-
Level 2
ity to safeguard equipment in- causes
tegrity becomes important as process
plants age, increase their capacities to Undersized Inappropriate
adjust to new market requirements, relief lines or relief line routing
equipment
undergo revamps or face new environ-
mental regulations.
Incorrect relief Block valves
In the past, approximately 30% of device set without involuntary
the chemical process industries’ (CPI) pressure closure prevention
losses could be attributed, at least in
part, to deficient relief systems [1]. Level 3
Furthermore, in an audit performed causes
refers to an installed relief device data, “the employer shall document 3. Overpressure scenario identifi-
with insufficient capacity to handle that equipment complies with rec- cation. In this step, the P&IDs are
the required relief load. The third ognized and generally accepted good examined in order to identify credible
type encompasses relief devices with engineering practices” [3] overpressure scenarios for each piece
incorrect set pressures, possibility 2. Process hazards analysis, which of equipment.
of involuntary blocking or hydraulic may include: What-if, hazard and op- 4. Overpressure scenario model-
problems. In addition to these prob- erability (HAZOP) study, failure mode ing. The fourth step is to model each
lems, other less frequent ones can be and effects analysis (FMEA), fault-tree credible overpressure scenario. Each
cataloged as miscellaneous deficien- analysis or equivalent methodologies. model is developed in accordance with
cies. A relief-system problem tree is In order to document that the plant the chosen reference standard (for
shown in Figure 1. equipment complies with recognized instance, API 520 and 521). The fol-
In a previous statistical analysis of and generally accepted good engineer- lowing calculations are typically per-
272 process units in the U.S., it was ing practices, the plant management formed during this step:
observed that [2]: must validate that the facilities are • Required relief load for each over-
• 15.1% of the facilities lacked relief de- protected against potential overpres- pressure scenario
vices on equipment with one or more sure scenarios, in accordance with • Required relief-device orifice area
potential overpressure scenarios accepted codes and standards, such for each overpressure scenario
• 8.6% of the relief devices were un- as API standards 520 and 521. An ef- • Relief line’s hydraulics
dersized fective relief-system-validation study • Knockout drum (KOD) liquid-sepa-
• 22% of the relief devices were im- comprises the following steps: ration verification
properly installed 1. Plant documents and draw- • Flare or vent radiation, dispersion
ings gathering. The first step in- and noise level calculations
Identifying potential problems volves obtaining and classifying the The overpressure scenario modeling
There are work methodologies that existing plant documents and draw- can be done in different ways, be it
allow identifying potential problems ings: process flow diagrams (PFDs), by hand calculations, spreadsheets
in relief systems. OSHA regulation mass and energy balances, product or by the use of steady-state or dy-
29 CFR 1910.119 is based on safety compositions, equipment and instru- namic relief-system simulation soft-
audits that use techniques such as ment datasheets, P&IDs, relief device ware. The results of the models are
process hazard analyses performed at datasheets, relief loads summaries, re- analyzed to identify potential prob-
regular intervals. The work methodol- lief line isometrics, one-line diagrams, lems. Table 1 summarizes the pos-
ogy established by this regulation to unit plot plan, and so on. sible relief system problems and the
identify safety hazards comprises two 2. Plant survey. The second step con- ways to identify them on the calcula-
basic steps [3]: sists of inspecting the installed relief tion results.
1. Process safety data gathering, devices to verify that they are free of
which includes the following: mechanical problems, to update and Available solutions
• Process chemical safety data fill-out missing data in the plant docu- There are various solutions for each
• Process technology data ments and to verify consistency be- type of relief system problem. The
• Process equipment data [materials tween the documents and drawings available solutions can be classified
of construction (MOCs), piping and and the actual as-built plant. During as: (a) modification of existing relief
instrumentation diagrams (P&IDs), plant surveys, other typical indica- system components, (b) replacement
design standards and codes, design tions of relief system problems are the of existing relief system components,
and basis of design of the relief sys- presence of pockets, leaks or freezing (c) installation of new relief system
tems, among others]. As part of these in relief lines and headers. components, or (d) increasing the reli-
Chemical Engineering www.che.com July 2010 39
Table 2. Conditions that increase the probability
and impact of relief system failure
Conditions that increase the probability of Conditions that increase the impact of
Feature Report relief system failure relief system failure
The plant has over 20 years of service The plant handles toxic, hazardous or
flammable fluids
ability of the emergency shut- The plant currently handles different prod- The plant handles gases
ucts to those it was originally designed for
down systems.
The modification of exist- The plant operates at a different load or The plant operates at high pressures
at different conditions to those it was origi-
ing relief-system components nally designed for
includes changes made to in- There have been contingencies that have The plant operates at high temperatures
stalled components, without required the replacement of equipment or
requiring their replacement. lines in the past
Some examples of this type of Rotating equipment (pumps, compressors) The plant has furnaces, or equipment
solution include the following: has been modified (for instance, new im- that adds considerable heat input to the
1. Recalibrating the pressure pellers) or replaced fluids
relief valve by readjusting The relief valves have not been checked or The plant has high-volume equipment
the set pressure (solution validated in the last ten years (such as columns, furnaces)
to incorrect set pressure) or Modifications have been made to existing The plant has exothermic reactors, or
relief valve lines (that is, they have been chemicals that could react exothermi-
the blowdown (solution to rerouted) cally in storage
inlet-line friction losses be- A complete and up-to-date relief valve in- The plant has large relief valves (8T10),
tween 3% and 6% of the set ventory is not available or the relief header has a large diameter
pressure) The relief load summary has not been up- The plant has a high number of opera-
2. Adding locks to relief lines’ dated in the last ten years tions personnel
block valves (to prevent in- A relief header backpressure profile is not The plant is located near populated
voluntary valve closure) available, or the existing model has not areas
The replacement of existing been updated in the last ten years
relief system components in-
volves substituting inadequate relief in which redundant instrumentation Deficiency No. 2
system elements for newer, appropri- and emergency shutdown valves are This type of deficiency involves under-
ate ones. Some examples of this solu- installed in order to cutoff the over- sized relief devices that are present on
tion are the following: pressure sources during a contingency. equipment with one or more potential
1. Replacing the installed pressure The main advantage of this type of overpressure scenarios.
relief valve, either for one with a solution is that it can significantly re- Case 2: Insufficient orifice area
larger orifice area (solution to un- duce the required relief loads, hence after changes in the stream com-
dersized relief device) or for one of a posing an economical alternative to position. In a petroleum refinery, a
different type (solution to excessive the installation of new relief headers, desalter that was originally designed
backpressure) knockout drums or flares. to process heavy crude oil was pro-
2. Replacing relief line sections to tected against a potential blocked
solve hydraulic problems, such as: examples of problems outlet by a relief valve on the crude
excessive relief-valve inlet-line fric- in aging systems outlet. When the refinery started
tion losses, excessive backpressure, What follows are examples of some processing lighter crude, simulations
excessive fluid velocity, pockets, typical relief-system problems that showed partial vaporization in the
among others can be found in aging process facilities relief valve. The vapor reduced the
The installation of new relief system and the recommended remedy. PSV capacity until it was insufficient
components involves the addition of to handle the required relief load. In
relief system elements that were not Deficiency No. 1 this case, the recommendation was to
included in the original design, such The first type of deficiency is when no replace the original PSV for one with
as the following: relief device is present on equipment a larger orifice and appropriate relief
1. New pressure relief valves, either with one or more potential overpres- lines.
on equipment lacking overpressure sure scenarios.
protection, or as supplementary Case 1. New overpressure scenario Deficiency No. 3
valves on equipment with under- after pump replacement. In a pro- The third type of deficiency involves
sized relief valves cess unit, a centrifugal pump was re- improperly installed pressure relief
2. New headers, knockout drums or placed for another one with a higher devices.
flares, when the revised relief loads head, without considering the down- Case 3: Excessive backpressure due
exceed the existing relief system stream system’s maximum-available to discharge line modifications. An
capacity, or when relief system seg- working pressure (MAWP). Since the existing vacuum-distillation column’s
regation (that is, acid flare/sweet downstream system was designed at PSV outlet lines were rerouted from the
flare, high-pressure/low-pressure the previous pump’s shutoff pressure, atmosphere to an existing flare header
flare) is required the installation of a higher shutoff due to new environmental regulations.
Increasing the reliability of the emer- pressure pump created a new blocked The installed PSVs were a conven-
gency shutdown systems is typically outlet scenario. Therefore, the instal- tional type, so with the new outlet-line
done via implementation of high in- lation of a new pressure safety valve routing, the backpressure exceeded
tegrity protection systems (HIPS), (PSV) was recommended. the allowable limit. A recommendation
40 Chemical Engineering www.che.com July 2010
Incorrect Correct
LO CSO
LO CSO
No measures are taken The block valves on PSV lines are kept
to prevent involuntary open via locks (LO) or car seals (CSO)
PSV blocking
Incorrect Correct
A PSV installed over the mist A PSV installed below the mist
eliminator is ineffective when eliminator is effective even if
the latter gets clogged the latter is clogged
Figure 2. The risk of blocking in a pressure safety valve Figure 3. Non-free-draining lines in installed relief lines, such
(PSV) can sometimes be readily identified on P&IDs as shown in these two constructions, may cause accumulation
of liquids that can hamper relief valve performance
was made to replace the existing PSVs tations, a non-ideal line arrangement only for a fire, additional discharges
for balanced bellows PSVs. is installed, creating pockets on relief were not considered by the designer.
Case 4: Incorrect PSV set pressure lines that may cause liquid accumu- However, the power failure also af-
due to static pressure differential. lation and hamper relief valve perfor- fected these drums. When this case
A liquid-full vessel’s relief valve was mance (Figure 3). was evaluated, the backpressure was
set to the vessel’s MAWP; however, the too high for the installed PSVs, so they
relief valve was installed several feet Deficiency No. 4 had to be replaced by piloted valves.
above the equipment’s top-tangent The fourth category of deficiencies is a
line. The static pressure differential miscellaneous grouping. MAINTAINING PROBLEM-
was such that the pressure inside the Case 8: Problems in an existing FREE RELIEF SYSTEMS
vessel exceeded the maximum-allow- flare network due to additional Some practical guidelines are offered
able accumulated pressure before the discharges. The additional dis- below to help the plant management
PSV would open. The problem was charges of various distillation-column to assess, identify and troubleshoot re-
solved by modifying the existing PSV, relief valves were rerouted to an exist- lief system problems.
recalibrating it to the vessel MAWP ing flare network because of new envi-
minus the static pressure differential. ronmental regulations. The additional Tip No. 1: Assess the risk
Case 5: Incorrect PSV set pressure discharges exceeded the system capac- Some factors tend to increase the prob-
due to higher operating tempera- ity, and the entire flare network and ability and impact of a relief system
ture. The temperature of a stream was emergency shutdown system had to be failure. Table 2 qualitatively shows
increased with the addition of new heat redesigned by selecting the optimum some of them. If several of the condi-
exchangers, and no attention was paid tie-in locations for the discharges, and tions shown on Table 2 apply, then the
to the set pressure of the thermal relief by implementing HIPS in order to re- plant management should consider
valve in the line. By increasing the tem- duce the required relief loads. planning a detailed study, such as a
perature, the pipe MAWP was reduced. Case 9: Sweet and sour flare mix- quantitative risk analysis (QRA), or a
The PSV set pressure was lowered to ing. When revamping a section of a relief-system validation study.
the new MAWP at the new working process unit’s relief headers, some acid
temperature plus a design margin. discharges were temporarily routed Tip No. 2: Maintaining up-to-
Case 6: Risk of blocking the relief to the sweet flare header in order to date relief-valve information
valve. A relief valve can be blocked for maintain operations. Soon afterwards, The plant management should main-
various reasons. Some of the most com- the header backpressure started to in- tain accurate, up-to-date relief-valve
mon include the lack of locked-open crease and scaling was detected upon data for maintenance and future ref-
(LO) or car-seal-open (CSO) indica- inspection. The acid gases could also erence. The following documents are of
tions in the PSV inlet- and outlet-line generate corrosion, as the sweet flare particular interest: (a) relief valve in-
block valves, and installing the PSV header material was inadequate to ventory, (b) relief loads summary and
above the mist eliminator on a sepa- handle them. (c) relief header backpressure profile.
rator. Both deficiencies can be readily Case 10: High- and low-pressure Relief valve inventory. The relief
identified on P&IDs (Figure 2). flare mixing. The discharges of low valve inventory is a list that contains
Case 7: Pockets. Relief lines going pressure PSVs located on drums were basic information and status for each
to closed systems should be self- routed to the closest flare header, which relief valve, which should include the
draining. It is not uncommon during was a high pressure header. Since the following:
construction that, due to space limi- design case for relief of the drums was • Valve tag
Chemical Engineering www.che.com July 2010 41
Feature Report
Table 3. Relief system validation study typical execution phases and deliverables
Phase Deliverable Deliverable description
Survey and Updated relief de- A list containing up-to-date, accurate data for each relief device located in the
data gathering vice inventory plant. The minimum data to be included on the list are as shown in Tip No. 1, and
they should be obtained by combining relief-valve manufacturer documentation
with onsite inspections
Updated P&IDs P&IDs showing the existing installed relief-device information: connection diam-
eters, orifice letter, set pressure, inlet- and outlet-line diameters and block valves
List of pockets A document identifying pockets on relief lines, with the appropriate photographs
Existing relief Updated relief A list containing the required relief loads for each applicable overpressure sce-
system model- loads summary nario of each relief device, the required orifice area and the relieving fluid prop-
ing erties, based on actual process information
Updated relief- A document showing a general arrangement of the relief headers and subhead-
network back- ers, along with updated backpressure profiles for the major plant contingencies
pressure profile
Updated relief de- A document containing the calculations for each relief device under actual op-
vice calculations erating conditions
List of relief sys- A document listing all of the deficiencies found in the existing relief system, cat-
tem deficiencies egorized by type
Relief system Conceptual engi- A document defining the modifications required to solve the relief system defi-
troubleshooting neering ciencies
• Process unit and area for each phase are described in Table be available to provide technical in-
• Location 3. If the plant management has spe- formation and verify the validity of
• Discharge location cific document formats, it should the consulting firm’s calculations. The
• Connection sizes provide them as part of the deliver- typical information that the consult-
• Connection rating able description. ing firm will request in order to com-
• Orifice letter The study may require a number of plete the study includes: relief device
• Manufacturer resources that are not readily avail- inventory, relief loads summary, relief
• Model able in the plant. If the plant man- device datasheets, mass and energy
• Type (conversion, ball, pilot) agement has available resources but balances, PFDs, P&IDs, equipment
• Set pressure lacks specialized software licenses, datasheets and relief line isometrics
• Allowable overpressure then it can assign some of the tasks for each evaluated process unit/area.
• Design case to inner resources, for example, survey The consulting firm may also request
• Installation date and data gathering. Tasks requiring process simulations, if available.
• Last inspection date expertise or software packages above
• Last calculation date the plant’s capabilities, such as com- Tip No. 4: When modeling, go
Relief loads summary. The relief plex distillation column, reactor sys- from simple to complex
loads summary contains all the over- tem or dynamic simulations, should Replacing a relief valve or header
pressure scenarios and relief loads be outsourced. section generates labor, materials, in-
for each relief device at the plant. The A consulting firm should be se- stallation and loss of production costs
data in this document can be used to lected based on its experience in that can only be justified when the
identify the critical overpressure sce- similar projects, technological ca- results of an accurate model identify
narios in the plant. pabilities (specialized software li- the need for it. However, developing
Relief-header backpressure pro- censes) and a reasonable cost esti- an accurate model for every relief de-
file. A backpressure profile of the en- mate. In order for the consulting firm vice in the plant can be impractical
tire relief network is valuable when to deliver an accurate estimate, the and costly, especially if only a small
evaluating the critical contingencies plant management should provide number of relief devices require re-
in the system, as it can be used to the scope definition along with suf- placement at the end.
identify relief valves operating above ficient information to identify each A practical compromise is to verify
their backpressure limits. relief device within the scope of the each system starting from a simple
project, its location and the possible model with conservative assump-
Tip No. 3: Planning and overpressure scenarios. These data tions, and developing a more accu-
executing a relief system study are available in the relief loads sum- rate model for those items that do
The execution of a typical, relief-sys- mary and relief device inventory. not comply with the required param-
tem validation study comprises three One person should be assigned on eters under such assumptions. This
phases: (a) survey and information the plant management side to manage approach minimizes the time and
gathering, (b) existing relief system the project, along with administrative effort dedicated to items, and con-
modeling and (c) relief system trou- personnel, and at least one in charge centrates on those items that could
bleshooting. The typical deliverables of technical issues; the latter should present problems.
42 Chemical Engineering www.che.com July 2010
For instance, for a blocked outlet rally wish to implement the quickest, References
downstream of a centrifugal pump most practical and least costly one. 1. American Institute of Chemical Engineers,
“Emergency Relief System (ERS) Design
and control valve system, the sim- For instance, when a relief valve’s Using DIERs Technology”, New York, 1995.
plest model is to assume a relief load inlet losses are between 3 and 6% of 2. Berwanger, Patrick, others, Pressure-Relief
equal to the pump’s rated capacity. If the set pressure, the valve blowdown Systems: Your Work Isn’t Done Yet, www.hy-
drocarbononline.com, July 7th, 1999.
the relief-valve orifice area is insuffi- can be adjusted instead of replacing 3. Occupational Safety and Health Administra-
cient under the previous assumption, the entire valve inlet line. tion, 29 CFR 1910.119 “Process Safety Man-
agement of Highly Hazardous Chemicals”.
the next step would be to read the
required relief load from the pump Tip No. 6: What to do after
curve with the control valve’s rated validation and troubleshooting Author
discharge coefficient and the valve’s A routine revalidation of the relief Sebastiano Giardinella is
a process engineer at Inelec-
downstream pressure equal to the re- system’s correct operation not only tra S.A.C.A. (Av. Principal con
lief pressure, ignoring piping friction bring that security to the plant man- Av. De La Rotonda. Complejo
Business Park Torre Este
losses. If the orifice area still seems agement over the integrity of its fa- Piso 5, Costa Del Este. Pan-
insufficient, then a rigorous hydraulic cilities, but also to third parties, such amá. Phone: +507-340-4842;
Fax: +507-304-4801; Email:
calculation of the entire circuit should as occupational safety organizations sebastiano.giardinella@
inelectra.com). He has six
be performed to determine the re- and insurance companies. The cost of years’ work experience in
quired relief load. a relief valve study may very well be industrial projects with a
special focus in relief systems design and evalu-
paid with a reduction in the plant in- ation, equipment sizing, process simulation and
Tip No. 5: Evaluate various surance premium. Furthermore, the bids preparation. He has participated in several
relief system evaluation studies, revamps and
solutions to problems image of a company that worries over new designs. Giardinella graduated as Chemical
As was mentioned earlier, there are the safety of its employees and the Engineer, Summa Cum Laude, at Universidad
Simón Bolívar in Venezuela and holds an M.S.
multiple solutions that are possible environment constitutes an impor- degree in project management from Universidad
Latina de Panamá. He has taken part as speaker
for a single relief system problem, and tant intangible benefit. n or coauthor on international conferences and is
the plant management would natu- Edited by Gerald Ondrey affiliated to Colegio de Ingenieros de Venezuela.
S U
www. So g
Into lids
6
che. Liquid
com s
Focu
s on
Gas
New Detec
Engin tion
eering
Closed Mater
Liquid ials
Dispe
Findin nsing
g the
To Fit Right
the Ap Gloves
Facts plicat
at Yo ion
ur Fin
Vacu gertip
um Pu s:
mps
Flowm
eter
News
Augu
st
2008
More and more, business in the Chemical Process Industries (CPI) is not
Rever
se
PAG Osmo
E 34 sis
Heat
Transf
er
Fluids
Hydro
local, it’s global. To keep up with this rapidly evolving marketplace, you
car
Prope bon
rties
Focu
Filtra
s on
need a magazine that covers it all, not just one country or region, not just
one vertical market, but the whole CPI.
tion
Facts
at Yo
r Finge ur
Lesso rtips:
ns-Le
Syste arned Valve
ms s
Preven
ting
Cakin
g
inherent limitations of
-10
current overpressure-
k=1.1
Temperature drop, °F
-40
Aubry Shackelford
Inglenook Engineering
-50 k=1.3
Brian Pack
BP North America
-60
I
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
n designing and sizing relief- Mach
device and effluent-handling
systems, one commonly over- Figure 1. Temperature drop relative to stagnation as a result of flowing
looked aspect of the perfor-
mance is examining the potential for Current-practice limitations stream piping system using an efflu-
low temperatures that can cause the Common practices for pressure re- ent-handling hydraulic model.
components of the system to reach lief and effluent handling are found Third, the temperature estimates are
temperatures below their respective, in numerous references [1–5]. The typically made for isothermal pressure-
minimum-design metal tempera- processes for estimating a discharge drop equations and do not account for
tures (MDMT), which may result in temperature and performing the out- effects like retrograde condensation.
brittle fracture with subsequent loss let pressure-drop calculations in the Fourth, some simplifications of the
of containment. This article points pressure-relief-device discharge pip- calculations that are used for the pur-
out limitations of the typical over- ing are limited in their ability to ac- poses of estimating the outlet pres-
pressure-protection-analysis philos- curately predict flowing temperatures sure drop do not represent flashing
ophy, discusses common sources of for many situations. effects (for example, highly subcooled
low temperatures for further inves- First, the discharge calculations flashing liquids are often choked at
tigation, and addresses possible de- are quite often only performed for the the bubblepoint; therefore, the sizing
sign remedies for MDMT concerns. controlling contingency for which the of the valve may assume the backpres-
The primary objectives of a process pressure relief device was sized, which sure is at the bubblepoint).
engineering evaluation of an efflu- does not necessarily represent the Finally, the temperature estimates
ent handling system (such as a flare most likely cause of overpressure or tend to be based on either relieving
system) include ensuring that opera- the cause resulting in the lowest dis- temperatures or isenthalpic flashes
tion of the pressure relief devices dis- charge temperatures. from relief conditions, which do not
charging into the collection system Second, the outlet pressure-drop account for kinetic energy effects.
(flare headers, for example) is not ad- calculations for individual pressure These effects can be substantial if the
versely affected; and that the effluent relief valves consider the outlet dis- developed velocity in the outlet piping
handling equipment are properly de- charge piping and potentially exclude is high and can be compounded when
signed to perform safely. The results of the remaining, downstream piping there are multiple relief devices dis-
an overpressure-protection design are system. This practice can result in a charging simultaneously into a collec-
the primary input for this engineering temperature discontinuity between tion system, or when large diameter
evaluation; however, there are several the calculated discharge temperature differences exist between the tail-pipe
potential gaps in the ability of these for the individual relief device and and the main effluent header.
data to identify situations in which that calculated for the same section Temperature drop. Figure 1 shows
the MDMT may be exceeded. of piping considering the entire down- the temperature drop from the stagna-
Chemical Engineering www.che.com July 2012 45
5
NRU
He
6
N2
1 LP
HP Auto-
refrig-
eration
4
Dehydration
2
Refrig-
Sweetening
erant
Demethanizer
Sweeten-
3 ing
NGL
surge
Flash
tank
Deethanizer
Stabilizer
Propane 7
loading
1. Process upsets in the NRU can cause low flowing temperatures to 5. Valve specifications for appropriate MDMT are
propagate downstream to after the gas/gas exchanger, which is typically common issues
designed based on normal operating temperatures 6. Common discharge lines can cause increased velocities, further
2. Relief of unstabilized condensate, a flashing liquid reducing flowing temperatures. Also depressuring is an
3. Breakthrough of high pressure gas or flashing liquid into the flash tank isentropic process that results in low flowing temperatures
4. Tube rupture in the chillers cause relief of expanded gas that starts cold 7. Potential relief of flashing volatile liquids that can cool
and is cooled further by JT effects significantly based on flashing
Figure 2. Typical schematic of NGL processing facility showing common areas of potential MDMT issues
tion temperature (Tstagnation) caused pressure relief valves to leak, or the thus represent a more likely source of
by the kinetic energy developed during effects of other inputs to the effluent fluid input to the collection system.
adiabatic compressible flow of an ideal handling system (such as pressure In addition, these control valves are
gas as a function of the Mach number control valves, depressuring valves, not necessarily sized to handle the full
for ideal gases having different ideal- pump seals and manual purge valves). flow of an overpressure scenario, re-
gas specific-heat ratios (k) (see Ref. 6, A leaking pressure-relief valve is typi- sulting in flow from both the control
Equation 6–128). For the purposes of cally considered an operational and valve and the pressure relief valve,
illustrating the temperature drop, a mechanical issue, not a cause of over- thereby exacerbating velocity effects.
stagnation temperature of 0°F (460R) pressure that needs to be evaluated for Finally, depressuring is a dynamic
was chosen. the sizing of the pressure relief valve process, releasing fluids of different
It is useful to note that while a stag- or for the effects on the downstream pressures and temperatures as a func-
nation temperature of 0°F seems un- collection system; however, many of us tion of time. Considering the most
likely for many cases, this stagnation in the warm Gulf Coast region of the likely behavior of a depressuring sys-
temperature is established after the U.S. recognize an ice-covering as indic- tem to be an isentropic expansion of
fluid has been relieved into the collec- ative of a leaking valve, and the fluids the residing fluid, the inlet fluid tem-
tion system (in other words, after the used in the evaluation of the pressure- peratures can drop significantly as the
isentropic process of flowing through relief-device sizing may not be repre- depressuring progresses.
the pressure-relief-valve nozzle and sentative of the normal process fluid
the subsequent adiabatic process of (for example, the external fire case, Low temperatures
expanding from the nozzle throat to which is a common design basis). While the potential for low flowing
the total backpressure that results in Pressure control valves may also be temperatures falling below the MDMT
Joule-Thompson (JT) cooling, both of called upon to “relieve” fluids, yet are exists in a variety of processing facili-
which can result in significantly lower commonly not accounted for in over- ties, the issue is especially apparent in
stagnation temperatures of the fluid pressure protection studies based on natural-gas processing facilities where
entering into the discharge piping). the desire to not include the positive high pressure, low temperature, low-
Additional limitations. Additional response of control systems in prevent- molecular-weight gases and volatile
gaps in the overpressure protection ing overpressure. In actual situations, liquids are present.
analysis include the common practice the basic process-control systems are Design considerations. Based on
of not considering the potential for expected to function as intended, and recent evaluations of several natural-
46 Chemical Engineering www.che.com July 2012
gas processing facilities with ethane fluid is at elevated temperatures, yet tion of overfilling of vessels), mitiga-
recovery capabilities, the authors during process upsets may experience tion of relieving conditions causing
have identified several common areas significantly lower temperatures. the low temperature excursion via
of concern that may provide a start- Figure 2 provides an overview of safety instrumented systems (SIS),
ing point for other gas processors’ these common areas of concern related performing mechanical stress analy-
investigations into this aspect of col- to low flowing temperatures. NGL ses to establish a better estimate of
lection system design, as well as for and propane processing-and-storage the MDMT per ASME B31.3 (with re-
process piping. These areas include equipment are examples of commonly placement of components not covered
the following: multiple inputs (such overlooked systems that can achieve by stress analysis as needed), add-
as pressure relief devices or control low flowing-discharge temperatures. ing supplemental fluid (such as gas
valves) discharging into subheaders These equipment usually have pres- or methanol) to raise the stagnation
having diameters close in size to the sure relief devices that are sized based temperature, rerouting the discharge
individual discharge piping diameter; on an external fire case, yet also have to a different location (such as to the
flashing liquid relief (unstablized con- the potential for relieving the liquid atmosphere), or conducting Charpy
densate, natural gas liquids [NGL] or either due to blocked discharges, leak- testing on the piping in question to
liquid propane); internal-boundary- ing relief valves or depressuring. establish the actual MDMT.
failure cases (tube rupture, for ex- Alternative solutions. While the For potentially leaking pressure-
ample) in gas chillers; cryogenic drain design issues related to low flowing relief valves, the options also include
operations (such as draining expander temperatures can be dealt with by recognizing the additional conse-
casing for maintenance); pressure- specifying appropriate metallurgy, quences in a risk-based inspection
relief-device MDMT specifications not there are other alternatives for con- protocol, installing rupture disks, or
commensurate with discharge piping sideration. These alternatives can adding skin thermocouples and low
MDMT; and pressure relief devices or include identifying ways to elimi- temperature alarms on the discharge
vents on the outlet of cryogenic cold- nate the cause of overpressure in piping to notify personnel of leakage
box sections where the normal process the first place (for example, preven- before the MDMT is crossed.
BRAND PRODUCTS
MPLIAN
CO T
S
ARM
INT
ON
I
AL TRAF
Environmental Manager
receive full
Receive full access
access today
todaybyby
visiting www.omeda.com/cbm/facts
visiting http://store.che.com/product/
Zone 2
Vent
Outdoors Zone 0
Gasoline
storage
10 ft (3m) tank 10 ft (3m) Gasoline
without storage
floating tank
roof without
Circular dike floating
around tank roof
Class 1, division 2
FIGURE 1. Shown here is a typical example of a Class I hazardous area uti- FIGURE 2. The example hazardous area shown in Figure 1 is here classified ac-
lizing division methods of area classification cording to the zones
Robert Schosker stances, where there exists the principles for working or gathering
Pepperl+Fuchs risk of explosion or fire that can be information out of a hazardous area
E
caused by an electrical spark or hot remain the same. It’s just that today
xplosion protection is essen- surface, requires specifically defined we have multiple options. In order to
tial for many companies, and instrumentation located in a hazard- exercise those options, we must first
those companies have deci- ous location. It also requires that determine if the danger of an explo-
sion makers. But before any the interfacing signals coming from sion exists and how severe it may be.
decisions can be made, there are a hazardous location are unable to
some important factors one must create the necessary conditions to What is a hazardous area?
consider. These factors include what ignite and propagate an explosion. Hazardous areas are most frequently
is most efficient and economical, as This risk of explosion or fire has found in places where there is a pos-
well as knowing the basics of explo- been the limiting factor when using sibility of an emission of flammable
sion protection; so the decision mak- electrical instrumentation because gas or dust. A hazardous area can
ers are headed in the right direction. energy levels were such that the occur in normal operation, in the
We will highlight many of the different energy limitation to the hazardous event of a fault, or due to wear and
“things to know,” but first, let’s step location was difficult, if not impossi- tear of seals or other components.
back in time and take a look at the ble, to obtain. For this reason, those Now the risk of an ignition of an
background of explosion protection. parts of the process that were con- air/gas mixture in this hazardous
sidered risky were controlled with area depends on the probability of
Backdrop pneumatic instrumentation. the simultaneous presence of the
After World War II, the increased following two conditions:
use of petroleum and its derivatives Moving forward • Formation of flammable or explo-
brought the construction of a great Now let’s move forward 70 years, sive vapors, liquids or gases, or
number of plants for extraction, refin- where almost everything you can combustible dusts or fibers with
ing and transformation of the chemi- think of can be found at the touch of a atmosphere or accumulation of
cal substances needed for techno- finger. From pneumatics to quad core explosive or flammable material
logical and industrial development. processors, information gathering • Presence of an energy source (elec-
The treatment of dangerous sub- has definitely changed, but the same trical spark, arc or surface temper-
50 Chemical Engineering www.chemengonline.com july 2015
Table 1. Defining areas for Divisions Table 2. The breakdown of Classes into subgroups
Class Type of Material Class Subgroup Atmospheres
Class I Locations containing flammable Class I Group A Atmospheres containing acetylene
gases, flammable liquid-produced va- Group B Atmospheres containing hydrogen and flammable process gases with
pors, or combustible liquid-produced more than 30 vol.% H2, or gases or vapors posing a similar risk level,
vapors such as butadiene and ethylene oxide
Class II Locations containing combustible Group C Atmospheres such as ether, ethylene or gases or vapors posing a similar
dusts risk
Class III Locations containing fibers and Group D Atmospheres such as acetone, ammonia, benzene, butane, cyclopropane,
flyings ethanol, gasoline, hexane, methanol, methane, natural gas, naphtha, pro-
pane or gases or vapors posing a similar threat
ature) that is capable of igniting the Class II Group E Atmospheres containing combustible metal dusts, including aluminum,
explosive atmosphere present magnesium, and their commercial alloys, or other combustible dusts
Determining hazardous areas in whose particle size, abrasiveness and conductivity present similar haz-
ards in the use of electronic equipment
a plant is normally performed by
experts from various disciplines. It Group F Atmospheres containing combustible carbonaceous dusts, including
carbon black, charcoal, coal or coke dusts that have more than 8% total
may be necessary for chemists, pro- entrapped volatiles, or dusts that have been sensitized by other materials
cess technologists, and mechani- so that they present an explosion hazard
cal engineers to cooperate with an Group G Atmospheres containing combustible dusts not included in Group E or
explosion-protection expert in order Group F, including flour, grain, wood, plastic and chemicals
to evaluate all hazards. The possible
presence of a potentially explosive Table 3. The Division Method
atmosphere as well as its properties Division Class I Class II Class III
and the duration of its occurrence
(gases and vapors) (flammable dust or powder) (flammable fibers or sus-
must be established. pended particles)
Also understanding terms such as In accordance with NEC In accordance with NEC
minimum ignition energy (MIE), upper 500.5 and CEC J18-004 500.6 and CEC 18-008 In accordance with NEC
and lower explosive limit (UEL/LEL), 500.5 and CEC 18-010
flash point, and ignition temperature Division 1 Areas containing dangerous Areas containing dangerous Areas containing dangerous
in the evaluation of your hazardous concentrations of flam- concentrations of flammable concentrations of flammable
mable gases, vapors or mist dusts continuously or oc- fibers or suspended particles
area will also provide a clearer direc- continuously or occasionally casionally under normal continuously or occasion-
tion on how severe a hazardous area under normal operating operating conditions ally under normal operating
might be. conditions conditions
In any situation involving an ex- Division 2 Areas probably not contain- Areas probably not contain- Areas probably not contain-
plosive material, the risk of ignition ing dangerous concentra- ing dangerous concentra- ing dangerous concentra-
must be taken into account. In addi- tions of flammable gases, tions of flammable dusts tions of flammable fibers or
tion to the nominal rating of materi- vapors or mist under normal under normal operating suspended particles under
operating conditions conditions normal operating conditions
als under consideration, parameters
related to the process involved are and utilized worldwide. may exist due to flammable gases,
especially important in the evalua- In rather simple terms, we can dif- vapors, or liquids within a plant or
tion. For example, the risk of explo- ferentiate between the International working environment. These are ex-
sion may be caused by the evapora- Electrotechnical Commission (IEC; plained by defining the type of haz-
tion of a liquid or by the presence of Geneva, Switzerland) (zones) and ardous material present, severity of
liquid sprayed under high pressure. the North American (division) proce- the hazard, and probability of the
It is also important to know which dures. The differences lie in the cat- hazard. It may also depend on the
atmospheric conditions are present egorization of hazardous areas, the likelihood of the hazard, risk of an
normally and abnormally. The range design of apparatus, and the installa- explosion, and the boundaries of the
of concentration between the ex- tion technology of electrical systems. hazardous location.
plosion limits generally increases as The categorization of these areas is This is usually determined by a
the pressure and temperature of the carried out in North America in ac- HAZOP (hazard and operability)
mixture increases. cordance with the National Electrical study and documented on a set
Code (NEC) NFPA 70, article 500. of electrical plot plans on record in
Divisions and zones The European Zone practice is de- every plant.
Once it has been determined that a scribed in IEC/EN 60079-10. For divisions, the type of material
hazardous area exists, it now needs So how does each work? First let’s is given by a class designation, as
to be classified. While the physical start at the basics, and then we’ll shown in Table 1. These can be bro-
principles of explosion protection cover each individually. ken down further into sub-groups,
are the same worldwide and are as shown in Table 2.
not differentiated, there are two dif- Defining the area Once we have determined the haz-
ferent and distinct models to define Hazardous location or area clas- ardous material we are working with,
your hazardous area —divisions and sification methods specify the dan- the probability of an explosion and
zones — both of which are accepted ger of fire or explosion hazards that boundaries must also be taken in to
Io L R
Interstice
U
C
Length of
S Uo
junction
P
FIGURE 3. Explosion-proof protection is based on FIGURE 4. In purging or pressurization protec- FIGURE 5. Intrinsic safety is based on the prin-
the explosion-containment concept, whereby the tion, a dangerous air/gas mixtures is not allowd ciple of preventing an effective source of ignition
enclosure is built to resist the excess pressure to penetrate the enclosure containing the electri-
created by an internal explosion cal parts that can generate sparks or dangerous
temperatures
ated, T1 to T6. The classes T2, T3
and T4 are divided into further sub-
consideration. The division method better rationalized by looking at the classes, as indicated in Table 7.
is divided into two areas: Division 1 example shown in Figure 2. In Europe, the apparatus are
and Division 2 (Table 3). These were With a slightly different approach, certified on the basis of design
created in 1947 when the NEC first IEC 600079-0 requires apparatus to and construction characteristics.
recognized that different levels of risk be subdivided into two groups, as From a practical point of view, the
exist in hazardous locations. Figure 1 shown in Table 5. two systems are equivalent, even
shows a typical example of a Class The groups indicate the types if there are minor differences, but
I hazardous area utilizing Division of danger for which the apparatus before you run out and choose the
methods of area classification. has been designed. Group I is in- most convenient method for you, it
In comparison to the division- tended for mines. Group II concerns is important that you consult your
based area classification, which is above-ground industries (electrical local authority having jurisdiction to
prevalent throughout North America, apparatus for hazardous areas with learn what method is allowed or, in
the zone-based architecture prevails potentially explosive gas (dust) at- fact, preferred.
in the rest of the world. mosphere except firedamp hazard- The initial steps to determine
Zones are similar in nature to divi- ous mining areas) and is subdivided whether a hazardous area exists and
sions where type of hazardous ma- into II G (gases) and II D (dusts). classify that area may seem rudimen-
terial present, severity of the hazard, Similar to divisions, the zones tary to some, but they are important
and probability of the hazard and offer a sub material classification as they now open up the multiple
boundaries must be determined. as well. Table 6 shows how this methods of protection, which may
Zones are in accordance with IEC/ approach compares to the North or may not be allowed, depending
EN 60079-10, which states that any American equivalent. on whether you classified your area
area in which there is a probability of Finally, when classifying your haz- by divisions or zones.
a flammable gas or dispersed dust ardous area, whether it be division
must be classified into one of the or zones, you must also classify the Protection methods
areas shown in Table 4. maximum surface temperature that There are three basic methods of
Similar to the division method of can go in to the hazardous area. The protection — explosion containment,
area classification, zones can be maximum surface temperature must segregation and prevention.
be below the minimum ignition tem- Explosion containment. This is the
Table 4. Defining areas by Zones
perature of the gas/dust present. only method that allows the explosion
Zone Type of material In North America as in Europe, six to occur, but confines it to a well-de-
Zone 0 An area in which an explosive air/gas temperature classes are differenti- fined area, thus avoiding the propaga-
mixture is continuously present or tion to the surrounding atmosphere.
present for long periods of time
Table 5. Apparatus Groups per IEC Flameproof and explosion-proof en-
Zone 1 An area in which an explosive air/gas 600079-0 closures are based on this method.
mixture is likely to occur in normal
operation Group Apparatus Segregation. This method attempts
Group I Apparatus to be used in mines where to physically separate or isolate the
Zone 2 An area in which an explosive air/gas
mixture is unlikely to occur; but if it the danger is represented by methane electrical parts or hot surfaces from
does, only for short periods of time gas and coal dust the explosive mixture. This method
Zone 20 An area in which a combustible dust Group II Apparatus to be used in surface includes various techniques, such
cloud is part of the air permanently, industries where the danger is rep- as pressurization, encapsulation,
over long periods of time or frequently resented by gas and vapor that has and so on.
been subdivided into three groups:
Zone 21 An area in which a combustible dust A, B and C. These subdivisions are
Prevention. Prevention limits the
cloud in air is likely to occur in normal based on the maximum experimental energy, both electrical and thermal,
operation safe gap (MESG) for an explosion- to safe levels under both normal op-
Zone 22 An area in which a combustible dust proof enclosure or the minimum igni- eration and fault conditions. Intrinsic
cloud in air may occur briefly or dur- tion current (MIC) for intrinsically safe safety is the most representative
ing abnormal operation electrical apparatus
technique of this method.
52 Chemical Engineering www.chemengonline.com july 2015
Table 6. Sub material classification for Zones Table 7. Temperature classes
Material Apparatus clasification Apparatus classification Ignition energy Tmax, °C Tmax, °F T Class in N.A.*
450 842 T1
Europe (*IEC) North America
300 572 T2
Methane Group I (mining) Class I, Group D
280 536 T2A
Acetylene Group IIC Class I, Group A > 20 µJ
260 500 T2B
Hydrogen Group IIC Class I, Group B > 20 µJ
230 446 T2C
Ethylene Group IIB Class I, Group C > 60 µJ
215 419 T2D
Propane Group IIA Class I, Group D > 180 µJ
200 392 T3
Conductive dust (metal) Group IIIC* Class II, Group E
180 356 T3A
Non-conductive dust Group IIIB* Class II, Group F
(carbon) 165 329 T3B
Cereal/flour Group IIIB* Class II, Group G 160 320 T3C
Fibers/suspended Group IIIA* Class III 135 275 T4
particles 120 248 T4A
100 212 T5
My application requirements fication, application, and cost — in
85 185 T6
Now the questions really start racing helping users find the ideal solution
in: Which should I use? Which one to match their needs. *N.A. = North America
offers the best protection? What if In addition to considering the nor-
all of my equipment is not low pow- mal functioning of the apparatus, area classification, there are many
ered? My plant is already using a eventual malfunctioning of the appa- other options when an area is classi-
technique; can I use another pro- ratus due to faulty components must fied using zones, but for now we will
tection method? Can they co-exist? be a consideration. And finally, all concentrate on the above as they
Who makes that decision? Why those conditions that can acciden- are most commonly used.
should I use one method over the tally occur, such as a short circuit, XP. The explosion-proof protection
other? Can I use two methods at open circuit, grounding and errone- method is the only one based on the
the same time? So many questions, ous wiring of the connecting cables, explosion-containment concept. In
all of which are very important, and must be evaluated. The choice of a this case, the energy source is permit-
with a little understanding of your specific protection method depends ted to come in contact with the dan-
process, they will guide you to best on the degree of safety needed for gerous air/gas mixture. Consequently,
method(s) to use. the type of hazardous location con- the explosion is allowed to take place,
Hazardous-area protection method sidered in such a way as to have the but it must remain confined in an en-
selection depends on three impor- lowest probable degree of an even- closure built to resist the excess pres-
tant factors: (1) area classification, (2) tual simultaneous presence of an sure created by an internal explosion,
the application and (3) the cost of the adequate energy source and a dan- thus impeding the propagation to the
protection method solution. gerous concentration level of an air/ surrounding atmosphere.
Area. Area classification depends on gas mixture. The theory supporting this method
the type of hazardous substances None of the protection methods is that the resultant gas jet coming
used, operating temperature, and can provide absolute certainty of from the enclosure is cooled rap-
explosion risk due to how often the preventing an explosion. Statistically, idly through the enclosure’s heat
dangerous substance is present in the probabilities are so low that not conduction and the expansion and
the atmosphere and the boundary even one incident of an explosion dilution of the hot gas in the colder
of the substance from various parts has been verified when a standard- external atmosphere. This is only
of the process. Area classification ized protection method has been possible if the enclosure openings
is determined by either the division properly installed and maintained. or interstices have sufficiently small
method or zone method. The first precaution is to avoid dimensions (Figure 3).
Application. Application character- placing electrical apparatus in haz- In North America, a flameproof en-
istics also affect which protection ardous locations. When designing closure (in accordance with IEC) is, as
method is used. For example, some a plant or factory, this factor needs a rule, equated with the "flameproof"
methods are more appropriate for to be considered. Only when there is designation. In both considerations,
large equipment protection, while no alternative should this application the housing must be designed for
others are more appropriate for high- be allowed. a x1.5 explosion overpressure. The
power applications. North American version “Explosion
Cost. Cost is also an important fac- Choosing the best method proof” (XP) must withstand a maxi-
tor for many engineers. For example, After carefully considering the above, mum explosion overpressure of x4.
if their application requires Division we can look at three more popular Furthermore, in North America,
2 protection, they may not want to methods of protection, XP (explo- the installation regulations (NEC 500)
purchase more expensive equipment sion proof/flameproof), purging and specify the use of metal conduit for
rated for Division 1. For that reason, pressurization, and intrinsic safety. the field wiring installation. It is also
it is important to understand the in- Although these are the most com- assumed here that the air-gas mixture
terplay of all three factors — classi- monly used methods in the division can also be present within the con-
Honeywell
Mike Baldi
Figure 1. Expansion of the
Honeywell Process Industrial Internet of Things
Solutions (IIoT) and cloud storage offers
benefits, but raises security
concerns
In Brief
cyber threats and
consequences
Defense in Depth
Adapting to the
needs of Operational
Technology
Risk-analysis
solutions
Next-generation
Firewalls
Endpoint protection
looking to the
Future
I
ndustrial cybersecurity risks are widely has brought tremendous benefits, along
appreciated. In April, the deputy direc- with more complex security concerns. The
tor of the U.S. National Security Agency, same convergence, however, has allowed
Rich Ledgett, warned that industrial the industrial world to adopt cyberdefense
control systems (ICS) and other critical in- techniques that have been widely used in
frastructure assets remain vulnerable to IT. This article discusses several key cy-
attack (Figure 1). Robust cyberdefense of bersecurity IT tools that can help industrial
industrial facilities remains an ongoing chal- facilities establish a layered cybersecurity
lenge for the chemical process industries system for its operations.
(CPI). The convergence between the world
of information technology (IT) and the world Cyber threats and consequences
of operational technology, in which con- The Stuxnet worm, a computer virus that in-
trol systems for industrial facilities reside, famously affected Iran’s nuclear centrifuges,
36 Chemical Engineering www.chemengonline.com july 2016
Honeywell
and the damage due to a cyberattack of a
German steel mill reported in 2014 are evi-
dence that cyberattacks can have physical,
real-world impacts. But it is not necessary
to prompt an explosion to cause significant
disruption. A cyber attack on Ukraine’s elec-
tric power grid, and subsequent widespread
power failure last December, was evidence
of that.
As NSA’s Ledgett put it, “You don’t need
to cause physical harm to affect critical infra-
structure assets.”
Cybersecurity risks are not easily ad-
dressed, however. One challenge is the
increasing sophistication of attacks. The
German government report on the steel
mill incident, for example, noted that the at-
tackers demonstrated not only expertise in
conventional IT security, “but also detailed
technical knowledge of the industrial control opportunity to store and process the large Figure 2. A layered approach
systems and production processes used in volumes of data resulting from the prolifera- to cybersecurity, with several
types of different cyber-
the plant.” tion of connected devices, such as with the defenses should be the ob-
Moreover, once the tools and knowledge IIoT. Again, however, it introduces new con- jective of industrial control
to enable such attacks are developed, they nection and communication channels that systems
are often quickly commoditized and shared, would-be cyberattackers will try to exploit.
allowing others with fewer technical skills to
use them. Defense in depth
Another challenge, however, is simply the In fact, the security issues related to the IIoT
increasing vulnerabilities introduced by the and Cloud storage result from the longer-
growth of intelligent, connected devices in in- term challenges surrounding the conver-
dustrial control systems. As Chris Hankin, di- gence between the IT and operational tech-
rector of the Institute for Security Science and nology (OT) worlds. Open platforms and the
Technology (ISST) at Imperial College, Lon- proliferation of third-party and open-source
don (www.imperial.ac.uk/security-institute), software in industrial control systems has
remarked recently: “Almost every component long brought the power and efficiencies from
of such systems now has fully functional com- the enterprise side of the business to the
puting capability and most of the connections process side. But along with those benefits,
will now be Ethernet, Wi-Fi or will be using the convergence also brings associated se-
Internet protocol.” curity concerns.
The growth of the Internet of Things — To complicate matters, while the vulner-
and, more specifically the Industrial Internet abilities on both sides — enterprise and op-
of Things (IIoT), in particular — is adding to erations — may be similar, the solutions are
both the number of devices and their con- often not directly transferable. The priorities
nectivity. Today, the IT research and advisory of each are necessarily different: while confi-
company Gartner Inc. (Stamford, Conn.; dentiality can be prioritized in the enterprise;
www.gartner.com) estimates 6.4 billion con- availability and integrity must, for the most
nected devices are in use worldwide. By part, take priority on the OT side. In prac-
2020, it forecasts, that total will reach 20.8 tice, a security solution cannot be allowed
billion. Moreover, heavy industries such as to shutdown operator access to data or de-
utilities, oil and gas, and manufacturing are vices that are essential to the safe running of
among the leading users. Each device and the plant, even if the security of those data is
connection expands the possible attack sur- at risk of being compromised.
face for cyberattacks. ISST’s Hankin acknowledged this reality in
Closely connected to the increasing num- his speech: “While there has been a conver-
ber of connected devices is the growth of gence between the two worlds [IT and OT],
the network of remote computer servers particularly in the past five years, there are
casually known as the “Cloud,” which pro- major differences, such as the fact the in-
vides access to infinitely scalable computing dustrial control systems (ICS) tend to have
power and storage. The Cloud provides an to operate in a time-critical way; they have to
In Brief
Safety and security
standards
What requires
Protection?
Applying Safety
principles to
security
W
hen the Stuxnet computer username/password authentication for re-
worm attacked programmable mote access. The attack was discovered
logic controllers (PLCs) at Ira- only after the data transmission volume ex-
nian nuclear facilities running ceeded the company’s data plan.
an integrated system, centrifuges were Cyber-related safety risks do not neces-
commanded to literally rip themselves sarily result from criminal activity. During the
apart. This clear demonstration of the link commissioning of one plant, for example,
between cybersecurity and safe industrial the failure of engineering software during
operations was a worldwide wakeup call for the recompiling of the memory mapped
plant managers, IT and automation manag- input (MMI) following a plant shutdown led
ers, safety engineers and many others. to a situation in which an incorrect modifi-
Of course, smaller-scale attacks are much cation was loaded into an integrated safety
more likely, and they are happening. At one controller, and then activated.
plant, where system maintenance was car- These incidents demonstrate the need for
ried out remotely, a cyber attack from abroad specific IT security improvements, and at the
revealed the vulnerability of using simple same time, raise broader questions about
42 Chemical Engineering www.chemengonline.com July 2016
HIMA Americas
Applying safety principles Written for engineers by engineers, Chemical Engineering delivers solid
The IEC 61511 (safety) and IEC
engineering essentials and developing industry trends to keep its readers
62443 (security) standards coincide
on the demand for independent lay- abreast of everything they need to keep their facilities running smoothly.
ers of protection. Together, these
standards prescribe:
• Independence between control Missing archived issues or what to share
systems and safety systems
• Reduction of systematic errors Chemical Engineering with your colleagues?
• Separation of technical and man-
agement responsibility
Visit www.chemengonline.com/chemical-engineering-magazine
• Reducing common-cause errors for more information.
The standards also reinforce that 27584
The quality of engineering processes, tools and associated services may be even more
important to overall safety results than BCPS and SIS hardware.
and using these same components for cause for potential failures, which would
different tasks – may not be considered infect both the BPCS and SIS.
sufficient leveraging of the integrity level The engineering tool for safety sys-
of the solution. tems should overcome these issues by
These commonly used components, remaining independent (to the greatest
in order to comply with the initial risk extent reasonably possible) from the
reduction requirements, will need to hardware and software environment.
maintain a risk reduction of greater than This is accomplished by having the com-
1,000 by less than 10,000. Practically, plete functionality of the safety engineer-
this means SIL 4, which is currently an ing tool, running in a Windows software
unachievable level. environment, implemented in a way that
allows it to be independent from Win-
Engineering’s key role in security dows functions. This concept allows
The quality of engineering processes, maximum protection from errors and
tools and associated services may be creates a trusted set of engineering data
even more important to overall safety that can be used to program the SIS.
results than BCPS and SIS hardware. Nevertheless, the engineering tool
Proper engineering includes the fol- should allow integrated engineering by
lowing aspects: maintaining interfaces that permit au-
• Reducing complexity by splitting tasks tomated transfer of configuration data
into independent modules (tag-oriented information as well as log-
• Properly defining and verifying ic-oriented data) from third-party sys-
interfaces tems into the trusted set of engineering
• Testing each module intensively data used for programming the SIS.
• Maintain the “four-eyes” principle Furthermore, having the same engi-
when reviewing engineering docu- neers in charge of programming the DCS
ments and results of implementation and safety system ignores the proven
tasks, according to IEC 60158-1, benefits of the checks and balances of
paragraph 8.2.18 independent thinking. For this reason,
Application of this strategy requires IEC 61508 is setting recommendations
46 Chemical Engineering www.chemengonline.com July 2016
for the degree of independence of par- products. Additionally, it should feature
ties involved in design, implementation easy-to-use engineering tools with fully
and verification of the SIS. integrated configuration and program-
ming and diagnostic capabilities.
IT security recommendations Applying these recommendations and
Cybersecurity and plant safety are so adhering to international standards for
intertwined in the connected world of in- separate BPCS and SIS systems help
dustrial processes that an equal commit- plant operators meet their obligation to
ment to both is required to achieve the protect people, communities, the envi-
needed protection. Following the recom- ronment and their own financial security.
mended international standards for func- The good news is that hardware, soft-
tional safety for PLCs (IEC 61508), safety ware and expertise are available today
instrumented systems (IEC 61511) and to help operators meet their obligations
cybersecurity (IEC 62443) provides a for the full lifecycle of their plants. n
path to a safe, secure facility. Edited by Scott Jenkins
For the most robust security and
reduced safety risks, the author ad- Author
vocates the traditional approach of Peter Sieber is vice president for
standalone SIS and BPCS units — ide- global sales and regional development
for HIMA Paul Hildebrandt GmbH (Al-
ally from different vendors — versus an bert-Bassermann-Strasse 28, 68782
integrated BPCS/safety system from Bruehl, Germany, Phone +49-6202
the same vendor. 709-0, p.sieber@hima.com), a leading
specialist in safety automation systems.
For valid security and safety reasons, Sieber is participating in the ongoing
it is also good practice for companies to effort by the steering committees work-
ing on functional safety and IT security
consider an independent safety system standards, IEC 61508 and IEC 62443, respectively. He has
built on a proprietary operating system. been actively involved in the development of the definition of
Of course, such a system can and should both functional safety guidelines and IT security guidelines
for process automation applications.
be completely compatible with DCS
Instrumentation and
High
level
to design a well-instrumented
and controlled system, and to Rotary
Receiving bins with bin
vent filters
Air valve
optimize its conveying velocity inlet
Run or position light
Blower
Amrit Agarwal FIGURE 1. This figure is a schematic flow diagram of the
Consulting Engineer conveying system with run and position lights to show the
operating condition of each component of the system
D
ilute-phase pneumatic con- Process logic description not stop the blower and the solids
veying systems must be Feeding solids into a conveying line feed at the same time.
operated in a certain se- that does not have an airflow with When a conveying cycle has
quence and have sufficient sufficiently high conveying velocity been completed and the solids flow
instrumentation and operating con- will result in plugging of the line. into the conveying line has been
trols to assure reliable operation To prevent this, solids must be fed stopped, the blower motor must
and prevent problems. This article into the conveying line only after continue to run for at least a few
discusses two subjects that are im- the required airflow has been fully more minutes to ensure that all of
portant for successful dilute-phase established. This requirement is the solids that are still inside the
conveying. Discussed below are de- met by allowing the solids feeder conveying line have been conveyed
sign guidelines for instrumenta- to start only after the blower has to the destination bin. If these sol-
tion and controls that can prevent been running for at least five ids are allowed to remain in the
operating problems, such as pipe- minutes. To do this, the rotary- conveying line, they may plug the
line plugging, downtime, equipment valve motor should be interlocked line when the system is restarted.
failure, high power consumption, with the blower motor so that the These solids may also cause con-
product contamination and more. blower motor has run for five min- tamination if a different solid is
The article also provides a simple utes before the rotary-valve motor conveyed in the next cycle.
methodology for finding out if the can start. Solids feed must stop immediately
presently used conveying veloc- When the conveying system is if the normal operating pressure of
ity is too low or two high and for running, the rotary-valve motor the blower increases by 10% and
making the required changes in must stop immediately in the event continues to rise. This is because
this velocity. that the blower motor stops for any the pressure increase is most likely
The required instrumentation reason. If the rotary valve is not due to the conveying line starting to
depends on the degree of automa- stopped, solids feed will continue plug. If the ongoing feed stream is
tion that is necessary, and whether and will plug the pipeline below not stopped, the pressure will keep
the system is to be controlled lo- the feeder. To remove this plug, the increasing, making the plugging
cally or remotely. When manual pipeline will need to be opened. This situation worse.
control of the conveying system required control option is imple- After stopping the feed, the
is used, problems can arise, espe- mented by interlocking the rotary- blower is allowed to run for about
cially if the operators do not have a valve motor with the blower motor five minutes in an effort to flush the
thorough understanding of the de- so that the rotary-valve motor stops plug. If the plug does not flush out
sign and of the required operating when the blower motor stops. and the blower pressure remains
method of the conveying system, or Should the conveying system need high, the blower motor should be
if they do not pay close attention to to be stopped, certain steps must be stopped. The plug is then removed
day-to-day operation of the system. followed: The first step is to stop the by tapping the pipeline to find the
For conveying systems — where solids feed, after which the blower plug location and opening up the
even a single error can result in a is allowed to run until the convey- plugged section of the pipeline.
large financial loss — a well-instru- ing line is empty and the blower Solids feed must also be stopped
mented and automated control sys- discharge pressure has come down if the receiving bin or silo becomes
tem is highly recommended. to the empty-line pressure drop. Do full, as indicated by its high-level
54 Chemical Engineering www.che.com March 2014
light and alarm. If the feed is con- ing status of the blower motor and is located far away from the rotary
tinued, the bin will overfill and the the rotary-valve motor valve, a second pressure indicator
solids will back up into the convey- • Position lights to indicate the di- should be provided just upstream of
ing line, causing pluggage. vert or through position of the di- the rotary valve.
If a conveying line has diverter verter valves These two measurements will
valves, the position of the diverter • Position lights to indicate the low show the overall pressure being
valves must be set up in a “through” and high levels in the receiving provided by the blower, and the
mode or in a “divert” mode before bin or silos pressure drop in the conveying line.
starting the blower and the solids • Run lights to show the operating In addition to local pressure indica-
feed. If the destination bin or silo status of the bin vent filters/dust tors, these pressure measurements
is changed for the next conveying collectors should also be provided in the
cycle, the diverter valves position Figure 1 shows in one glance how control room using pressure
must be changed before the con- the conveying system has been set transmitters.
veying blower and the rotary valve up, and the operating status of all Digital pressure indicators are
are started. components of the system. better than the analog type, because
Graphic control panel. In the Monitoring conveying air pres- they can show the pressure much
central control room, a graphic sure. Conveying pressure is a key more accurately, up to two decimal
panel (Figure 1) should be pro- parameter in pneumatic convey- points. These pressure measure-
vided to show a schematic diagram ing systems. It must be regularly ments should be archived on the
of the conveying system, start- monitored from the control room computer so that historical data are
ing from the air supply blower as well as locally at the blower. available if needed in the future.
to the receiving bins or silos. For measurement of the conveying An alarm for high blower-discharge
This panel should have the follow- pressure, a locally mounted pres- pressure should also be provided in
ing lights: sure indicator should be provided at the control room.
• Run lights to indicate the opera- the blower discharge. If the blower Monitoring blower discharge air
0
BATCH 1 BATCH 2 BATCH 3
Circle 22 on p. 68 or go to adlinks.che.com/50974-22
Chemical Engineering www.che.com March 2014 55
Solids Processing
temperature. A locally mounted veying characteristics of solids rotary valve if the conveying pres-
temperature indicator should be being conveyed, it must be moni- sure starts to increase (indicating
provided at the blower discharge, tored closely. start of formation of a line plug); 2)
and also at the blower after-cooler Rotary-valve motor interlocks Allow the blower motor to continue
discharge if an air cooler is used. with the blower motor. A manu- to run for the selected time, such
This temperature is needed to ally adjustable timer with a selec- as 10 to 15 minutes (in an effort
carry out calculations for the “as- tor switch should be provided in to clear the line plug); and 3) Re-
built” conveying system. If this air the control room to provide three start the rotary-valve motor if the
temperature can affect the con- functions: 1) Automatically stop the conveying pressure falls to the nor-
mal pressure.
Diverter valves. Position lights are
provided in the control room graphic
panel to indicate if the valves are in
the “through” or “divert” position.
Receiving bins. Low- and high-
level lights are provided in the
graphic panel for the receiving bins.
An alarm should be provided in the
control room to indicate high level
in the bins. At the high level, the ro-
tary valve motor should be stopped
automatically.
Bin vent filters/dust collectors.
The bin vent filters or the dust col-
lectors on the bin vents must be
running before the conveying sys-
tem is started. A “run” light for the
Understanding Protective Coatings filter should be provided in the
graphic panel.
in Hot Environments Pressure drop indicators should
be installed locally to show the pres-
sure drop across the filter elements.
The use of fireproofing, high-temperature, and other coatings to Their locations should be easily ac-
protect infrastructure in high-temperature facilities is becoming more cessible to the operating staff. For
widespread around the world. The 2014 Bring on the Heat Conference conveying materials that have high
is focused on providing an informative look into how these coatings are dust loading, alarms for low- and
used and how they benefit different industries. This event will provide high-pressure drops should be pro-
presentations, case studies, and forum discussions on the following vided in the control room. The low-
topics: pressure drop alarm would indi-
cate a ruptured filter element, and
• Corrosion under insulation • Passive fire protection the high pressure drop alarm
• Thermal insulation coatings • Coatings needs for owners would indicate a completely clogged
• Thermal spray aluminum filter element.
Instrumentation checklist
Register Today A summary of the instrumentation
requirements, as described above, is
provided below:
Register by May 16 For the blower:
• Local and control room mounted
to SAVE! running lights for the blower
motor
• Local pressure indicator at the
blower discharge
To register or for more information go to • Local temperature indicator at
www.nace.org/both2014 the blower discharge
• Local temperature indicator at
the blower after-cooler discharge,
Circle 27 on p. 68 or go to adlinks.che.com/50974-27 for applications using a cooler
56 Chemical Engineering www.che.com March 2014
Saltation line, Dilute
joining pressure phase:
M Unstable minima and suspension
flow saltation velocity flow
d
H
be
∆P
L
Log (pressure drop per unit length), L
W2
ed
=
ck
d ing • Alarms for low- and high-pres-
oa
Pa
e
lid C
im
K So (optional)
reg
W 1
G g= Graphic control panel:
din
ble
loa • Graphic panel showing the con-
sta
G
l ids B
Dense Un So e veying system route with run
F ip
phase
ptyp lights for the blower motor and
regime E Em rotary valve motor, position lights
Dilute for the diverter valves, low- and
phase:
strand D high-level lights for the receiving
flow bins, and run lights for the bin-
A vent filters
Saltation velocity/minimum pressure
Kim VanCamp
Emerson Process
Management
In Brief
Alarm management
Performance Metrics
Alarm system
example metrics
Alarm priority
distribution
Alarm source
contribution
stale alarms
Closing remarks
Figure 1. A better understanding of alarm system metrics can lead to more focused remedial actions and help to make the
plant safer
D
o you routinely receive “alarm man- defining a valid alarm), alarm management
agement performance” reports, methods (for instance, rationalization tech-
or are you expected to monitor niques), justification (such as the benefits of
a managerial dashboard equiva- investing in alarm management) and tools
lent? What do you look for and what does it (including dynamic alarming enablers). This
mean? We all know that fewer alarms mean article provides a different perspective. Writ-
fewer operator interruptions and presumably ten for process plant operation managers or
fewer abnormal process or equipment con- others that routinely receive alarm manage-
ditions. But a deeper understanding of the ment performance reports, this article aims
more common alarm-management metrics to explain the most common metrics, with-
can yield greater insight, leading to more fo- out requiring an understanding of the alarm-
cused remedial actions and ultimately to a management discipline in depth.
safer, better performing plant (Figure 1).
This article reviews the now well estab- Alarm-management KPIs
lished benchmark metrics associated with The first widely circulated benchmark met-
the alarm-management discipline. Most arti- rics, or key performance indicators (KPIs), for
cles previously published on alarm manage- alarm management relevant to the chemical
ments cover alarm concepts (for example, process industries (CPI) were published in the
50 Chemical Engineering www.chemengonline.com march 2016
Table 1. Example of typical alarm performance metrics, targets and action limits
Metric Target Action limit
Average alarm rate per operator (alarms per day) < 288 > 432
Average alarm rate per operator (alarms per hour) < 12 > 18
Average alarm rate per operator (alarms per 10 minutes) 1–2 >3
Percent of 10-minute periods containing > 10 alarms < 1% > 5%
Maximum number of alarms in a 10 minute period ≤10 > 10
Percent of time the system is in flood < 1% > 5%
Annunciated priority distribution (low priority) ~80% < 50%
Annunciated priority distribution (medium priority) ~15% > 25%
Annunciated priority distribution (high priority) ~5% >15%
Percent contribution of top 10 most frequent alarms < 1% to ~5% > 20%
Quantity of chattering and fleeting alarms 0 >5
Stale alarms (number of alarms active for more than >24 hours) < 5 on any day >5
1999 edition of the Engineering Equipment publications state that an average long-term
and Materials Users Association publication rate of new alarms occurring at a frequency
EEMUA-191 Alarm Systems – A Guide to of up to 12 alarms per hour is the maximum
Design, Management and Procurement [1]. manageable for an operator. Suggested
Later works from standards organizations, performance levels such as this can pro-
such as the 2009 publication International So- vide a reasonable starting point if you are
ciety of Automation (ISA) 18.2 Management just beginning an alarm-management pro-
of Alarm Systems for the Process Industries gram. But before deciding what constitutes
[2] and the 2014 publication IEC62682 Man- a reasonable set of targets for your site, you
agement of alarms systems for the process should also consider other firsthand inputs,
industries [3], built upon EEMUA-191 and like surveying your operators and reviewing
have furthered alarm-management thought in-house studies of significant process dis-
and discipline. For example, they provide a turbances and alarm floods. Note that more
lifecycle framework for effectively manag- research into the human factors that affect
ing alarms and establish precise definitions operator performance is needed to validate
for core concepts and terminology. Yet fif- and potentially improve on the current pub-
teen years later, little has changed regarding lished performance targets. Important work
the metrics used to measure alarm-system in this area is ongoing at the Center for Op-
performance. This consistency in measure- erator Performance (Dayton, Ohio; www.
ment has been positive in many respects, operatorperformance.org).
leading to the wide availability of generally
consistent commercial alarm analytic re- Alarm system example metrics
porting products, from both control-system A typical alarm-performance report contains
vendors and from companies that specialize a table similar to Table 1, where the metrics
in alarm management. Consequently, selec- and targets are based upon, and in many
tion of an alarm-analysis product may be cases, copied directly from, the EEMUA-
based on factors such as ease of use, inte- 191, ISA-18.2 and IEC62682 publications. It
gration and migration, reporting capabilities, is also common to see locally specified ac-
price, support availability and so forth; with tion limits based on a site’s alarm philosophy.
reasonable certainty that the KPIs derived When a target or action limit is exceeded, it
from the chosen product can be interpreted is important to ask: what problems are likely
consistently and compared across sites and contributing to the need for action, and what
across differing process control, safety and are the actions? These questions are the
other open platform communications (OPC)- focus of the following discussion.
capable alarm-generating sources.
In addition to defining the KPI measure- Average alarm rate
ments, the EEMUA-191, ISA-18.2 and The average alarm rate is a straightforward
IEC62682 publications also suggest perfor- measure of the frequency with which new
mance targets, based in large part on the alarms are presented to the operator, ex-
practical experience of the companies par- pressed as an average count per day, hour
ticipating in the committees that contributed or per 10-minute interval. As alarm frequency
to each publication. As an example, these increases, an operator’s ability to respond
600
400 n Critical
n Warning
n Advisory
200
5/6/2009
5/7/2009
5/8/2009
5/9/2009
5/10/2009
5/11/2009
5/12/2009
5/13/2009
5/14/2009
5/15/2009
5/16/2009
5/17/2009
5/18/2009
5/19/2009
5/20/2009
5/21/2009
5/22/2009
5/23/2009
5/24/2009
5/25/2009
5/26/2009
5/27/2009
5/28/2009
5/29/2009
5/30/2009
5/31/2009
Date
correctly and in time to avoid the ultimate mal condition requiring a timely response
consequence of inaction decreases. If the is occurring
rate is excessively high, it is probable that • Chattering or other frequently occurring
some alarms will be missed altogether or the nuisance alarms are present. These often
operators will ignore them, thus eroding their originate from non-process alarm sources
overall sense of concern and urgency. So of marginal interest to the operator, such
clearly it is an important metric. as field devices or system hardware diag-
Averages can be misleading, however, be- nostics. Chattering alarms can also indi-
cause they provide no sense of the peaks in cate an incorrect alarm limit or deadband
the alarm rate, making it difficult to distinguish • Redundant alarms, where multiple alarms
“alarm floods” from steady-state “normal” are presented when a single abnormal situ-
operation. Consequently, most alarm per- ation occurs. An example is when a pump
formance reports supplement this basic KPI is shut down unexpectedly, generating a
value with a timeline view or separate calcu- pump fail alarm in addition to alarms for low
lation of alarm rates for both the times when outlet flow and low discharge pressure
operation is normal and for times of an alarm • A problem with the metric calculation is oc-
flood. Figure 2 presents a typical example. curring. A correct calculation only counts
The average alarm rate of 16.5 alarms per new alarms presented to the particular
hour exceeds the target KPI value of 12 from operator or operating position for which
Table 1, but is slightly less than the action the metric is intended, taking into consid-
limit of 18 per hour, and so might not raise eration any by-design threshold settings or
concern, while the timeline view shows that other authorized filtering mechanisms that
there are significant periods of time where cause fewer alarms to be presented to the
the performance is unacceptable. operator than may be recorded in system
Common contributors to an excessively event logs
high alarm rate include the following:
• The alarm system is being used to notify Peak alarm rate
the operator of events that do not consti- The two metrics — the percentage of
tute actual alarms, such as communicat- 10-minute periods with more than 10
ing informational “for your information” alarms, and the percent of time spent in
messages, prompts, reminders or alerts. an “alarm flood” state — are calculated dif-
According to ISA-18.2, an “alarm” is an in- ferently, but are highly similar in that they
dication to the operator that an equipment quantify how much of the operator’s time
malfunction, process deviation or abnor- is spent within the highly stressful circum-
stance of receiving more alarms than can
Figure 3. Pie charts can sup- New alarm activation rate distribution
plement alarm performance be managed effectively.
reports and give information 6.6%
n Acceptable EEMUA-191 defines the start of an alarm
on how much time is spent in 10.1% (0–1 per 10 min.) flood as a 10-minute period with more than
the acceptable range
n Manageable 10 new alarms, continuing through subse-
(2–4 per 10 min.) quent 10-minute intervals until reaching
n Demanding a 10-minute interval with fewer than five
20.0% 63.4% (5–9 per 10 min.) new alarms. Equally acceptable is to de-
n Unacceptable fine a flood simply as a 10-minute period
(≥10 per10 min.) with more than 10 new alarms. Often, an
alarm-performance report will supplement
these two metrics with a pie chart (Figure
52 Chemical Engineering www.chemengonline.com march 2016
3) that segments the report period into Alarm priority distribution Figure 4. When the number of high-pri-
10-minute periods that are catego- ority alarms exceeds that of low-priority
alarms, the methodology of how alarms
rized into named alarm-rate ranges, 8.7%
are assigned priority should be evaluated
such as acceptable, manageable, de-
manding and unacceptable. n Medium
Another commonly included metric in 51.8% n High
39.4%
the alarm-performance report, the peak n Low
number of alarms within a 10-minute
period, is a straightforward measure
of the degree of difficulty of the worst-
case alarm flood for the operator. In
poorly performing alarm systems, it is
common to see peak alarm counts in ties: low, medium, high and very-high.
a 10-minute period that exceed 250, There are a number of well accepted
a total that would overwhelm even the methods for assigning priority, the most
most highly skilled operator. common being a systematic guided
Common contributors to high peak- (selection-based) consideration of the
alarm-rate frequency and severity in- severity of the consequence of inaction
clude the following items: combined with the time available for the
• Multiple redundant alarms for the operator to take the required action.
same abnormal condition. The op- Conventional wisdom says that the an-
timum situation is of course that any nunciated alarm-priority distribution ex-
single abnormal event will produce perienced by the operator for low-, me-
just one alarm, representing the best dium- and high-priority alarms should
choice in terms of operator compre- be in an approximate ratio of 80, 15 and
hension and the quickest path to take 5%. Ultimately however, the goal should
remedial action. This requires study of be to guide the operator’s determination
alarm causes and often leads to the of the relative importance of one alarm
design of conditional, first-out or other compared to another, based on their
form of advanced alarming logic importance to the business.
• Cascading alarms. The sudden Figure 4 illustrates a situation where
shutdown of equipment often trig- the number of high-priority (critical)
gers automated actions of the control alarms being presented to the operator
system, which in turn, triggers more far exceeds the low-priority (advisory)
alarms alarms, suggesting the need to review
• False indications. When routine the consistency and methodology of
transitions between process states the priority assignment.
occur, the alarm system is not usually Common contributors to out-of-bal-
designed to “follow the process,” so ance alarm-priority distributions include
it can therefore produce a multitude the following:
of false indications of an abnormal • Alarm prioritization (a step in the ratio-
condition. Likewise, logic is typically nalization process) has not been per-
required to detect state changes and formed and alarm priorities have been
suppress or modify alarms accord- left at their default values
ingly • Misuse of the priority-setting scheme
Some systems provide specialized to classify alarms for reasons other
alarm views that present alarms in a than providing the operator with a tie-
graphical pattern to aid an operator’s breaker during alarm peaks. For ex-
comprehension of peak alarm events ample, using priority to classify alarms
and their associated causality, supple- by impact categories, such as envi-
menting the classic alarm list to help pro- ronmental, product quality, safety/
vide a built-in layer of defense against the health, or economic loss
overwhelming effects of an alarm flood. • Lack of discipline in setting prior-
ity based on consideration of direct
Alarm priority distribution (proximate) consequences rather than
When faced with multiple alarms, the ultimate (unmitigated) consequences.
operator must decide which to address While it may be the case that a de-
first. This is — or should be — the basis signed operator action could fail, fol-
for assigning priority to an alarm. Most lowed by a protective system failure,
systems will employ three or four priori- followed by a subsequent incorrect
IIUP16P1
FICUP1516
IIPX15P1
OPC_FI-N2-051
TIFH42106
FIC-1252
PICFP2043
TIFG41106
FIFC1054
n Alarms
— Cumulative % (deadband).
Stale alarms
Alarm source
A stale alarm is one that remains annunci-
ated for an extended period of time, most
Figure 5. A small number human response, such what-if consider- often specified as 24 hours. Stale alarms
of alarm sources can often ations are likely to lead to a vast skewing are surprisingly challenging to quantify.
account for the majority of
alarms of alarm priorities toward critical Metrics based on event histories require the
presence of both the start and ending alarm
Alarm source contribution event in order to compute an alarm’s an-
The percent of alarms coming from the top- nunciated duration. There is no event rep-
ten most frequent alarm sources relative to resenting the attainment of a certain age of
the total alarm count is a highly useful met- an annunciated alarm. Thus, it is common
ric for quantifying, identifying and ultimately to miss counting stale alarms if their acti-
weeding out nuisance alarms and alarm- vation event or all-clear event falls outside
system misuse. This is especially true if the the range of dates and times covered in the
alarm performance report covers a range of event history. Consequently, there are alter-
time where operations were routine and with- nate methods for quantifying stale alarms,
out significant process upsets or equipment such as periodic sampling of the active
failures. The top-ten alarm sources often alarm lists at each operator workstation, or
provide “low-hanging” fruit for alarm-man- simply counting the number of alarms that
agement performance improvement. They attained an age greater than the threshold
are a handful of alarms, which if addressed, age. Given this variation in methods, it is im-
will create a noticeable positive change for portant to exercise caution when compar-
the operator. ing stale-alarm metrics across different sites
Figure 5 shows a pattern observed in that may be using different alarm-analytic
many control systems, where as few as applications.
ten alarm sources (like a control module or In addition to being hard to quantify, stale
transmitter) out of the many thousands of alarms can also be some of the most difficult
defined alarm sources, collectively account nuisance alarms to eliminate. Thus in some
for about 80% of all of the alarms presented respects the upward or downward trend in
to the operator. In this example, the first stale alarm counts provides an informal in-
alarm source (FIST111) alone was respon- dication of the overall ongoing health of the
sible for 15% of all of the alarms presented alarm management program.
to the operator. Common contributors to stale alarm
Another related metric is the count of counts include the following:
chattering alarms — alarms that repeatedly • Routine transitions between process states
transition between the alarm state and the where the alarm system is not designed to
normal state in a short period of time. The adapt and therefore provides false indica-
specific criteria for identifying chattering tions of an abnormal condition
alarms vary. The most common method is • Alarms associated with standby or idle
to count alarms that activate three or more equipment
times within one minute. • Alarms configured to monitor conditions no
When the top-ten alarm sources generate longer relevant or available, an indicator of
over 20% of all the alarms presented to the poor management-of-change processes
operator, it is a strong indicator that one or • Alarms that are essentially latched due to
both of the following is the case: excessive application of hysteresis
• Some of those alarms are nuisance alarms • Alarms that persist beyond the called-for
54 Chemical Engineering www.chemengonline.com march 2016
operator action, waiting for maintenance methods for corrective action, one outstand-
action. This likely constitutes an incorrect ing resource is Ref. 4. n
use of the alarm system, using it as a re- Edited by Scott Jenkins
cording method for outstanding mainte-
nance actions References
In conjunction with reviewing the number 1. EEMUA Publication 191 — Alarm Systems: A Guide to Design,
of stale alarms or the list of stale alarms, it is Management and Procurement – Third edition, published by the
Engineering Equipment and Materials Users Association in 2013.
also important to review what alarms have
2. ANSI/ISA–18.2–2009 — Management of Alarm Systems for
been manually suppressed (thus removing the Process Industries – approved June 23, 2009. ISBN: 978-1-
them from the view of the operator). Sup- 936007-19-6.
pressing the alarm will remove a stale alarm 3. ANSI/ISA–18.2–2009 — Management of Alarm Systems for
the Process Industries – approved June 23, 2009. ISBN: 978-1-
from the alarm list (effectively reducing the 936007-19-6.
number of stale alarms), but will not address 4. International Society of Automation. Technical Report ISA-TRI
the underlying condition. 18.2.5, Alarm System Monitoring Assessment and Auditing, ISA.
2012.
Closing remarks
This article touches on just some of the key Author
alarm-system performance metrics and Kim VanCamp is the DeltaV marketing prod-
what the numbers represent, in terms of the uct manager for alarm management at Emer-
son Process Management (8000 Norman
issues that lay behind them and possible ac- Center Drive, Bloomington, MN 55437; Phone:
tions to address them. With this understand- 1-952-828-3500; Email: Kim.VanCamp@
ing, periodic reviews of alarm-performance emerson.com). He joined Emerson in 1976
and has held senior assignments in manufac-
reports should lead to more focused actions turing, technology, field service, customer
that can improve operator effectiveness and service, service marketing and product mar-
keting. VanCamp is a voting member of the
thereby reduce the risks for economic loss, ISA-18.2 committee on Management of Alarm Systems for the Pro-
environmental damage or unsafe situations. cess Industries and has published multiple papers on alarm man-
For further reading on these and other alarm agement. He holds a bachelor’s degree in electrical engineering
from the University of Nebraska.
performance metrics, including suggested
American Filtration and Separations Society • May 9-11, 2016 • Houston Marriott Westchase, Houston, TX
The American Filtration & Separations Society invites you to
Oil, Gas and Chemicals Filtration & Separations Conference – Expo
Conference website: http://spring.afssociety.org/, Contact Conference Chair for questions, David Engel +1 (832) 296-6624
Conference Features Technical Conference Topics Participating Companies
• 3 Plenary sessions • Filtration • Berry Plastics (formerly AVINTIV)
• 3 Concurrent tracks • Coalescing • Delta Pure Filtration
• 72 Technical papers • Adsorption & Absorption • Dorstener Wire Tech
• Student Poster Competition • Air/Gas Purification • GKD – USA Metal Fabrics
• Short Courses on Monday, May 9 • Bulk Separations and Cyclonics • IFTS Filter Testing
• Vendor Expo • Chemical Assisted Separations • Industrial Netting
Plenary Speakers • Equipment & Systems • Nexo Solutions
• Larry Ryan – Dow Chemical Company • Media Technology • Onyx Specialty Paper
• Michael Spearman – Otronix • Produced Water • Sefar Filtration
• Scott Northrop – Exxon Mobil • R & D Innovation • Spifil Filtration
Don’t miss the premier filtration & separations Expo in the USA.
Expo only passes are complimentary but registration in advance is required.
Circle 2 on p. 90 or go to adlinks.chemengonline.com/61493-02
Bill Hollifield
PAS Inc.
In Brief
alarm rates
Averages can be
misleading
Alarm rationalization
Alarm management
work processes
Concluding remarks
M
odern distributed control sys- Figure 1. Alarm rates on the order of thousands per day are
tems (DCS) and supervisory con- not uncommon in some CPI facilities
trol and data acquisition (SCADA) to create an alarm for any imaginable con-
systems are highly capable at dition. A simple analog sensor, such as one
controlling chemical processes. However, for temperature, will likely have a dozen alarm
when incorrectly configured, as is often the types available by simply clicking on check
case, they also excel at another task — gen- boxes in the device’s configuration. Without
erating alarms. It is common to find alarm following sound alarm-management prin-
rates that exceed thousands per day or per ciples, the typical results are over-alarming,
shift at some chemical process industries nuisance alarms, high alarm rates and an
(CPI) facilities (Figure 1). This is a far greater alarm system that acts as a nuisance distrac-
number than any human can possibly handle tion to the operator rather than a useful tool.
successfully. This article examines the nature Whenever the operators’ alarm-handling
of the problem and its cure. capacity is exceeded, then operators are
The alarm system acts as an intentional in- forced to ignore alarms, not because they
terruption to the operator. It must be reserved want to do so, but because they are not able
for items of importance and significance. An to handle the number of alarms. If this is the
alarm should be an indication of an abnormal case, the average, mean, median, standard
condition or a malfunction that requires op- deviation, or other key performance indica-
erator action to avoid a consequence. Most tors (KPIs; see Part 1, p. 50) for alarms do
alarm systems include interruptions that not matter, because plant managers have no
meet this definition, but also many miscel- assurance that operators are correctly ignor-
laneous status indications that do not. ing inconsequential alarms or are paying at-
A major reason for this situation is that con- tention to the ones that matter. This situation
trol system manufacturers make it very easy contributes to many major accidents.
56 Chemical Engineering www.chemengonline.com march 2016
Alarm rates 3,000 Annunciated alarms per day Peak rates
— Annunciated alarms
The International Society of Automation (ISA; — Annunciated alarms without the 10 most frequent 9,195
Research Triangle Park, N.C.; www.isa.org) 2,500 14,899
Standard 18.2 on alarm management identi-
fies the nature of the problem and offers a 2,000
variety of assessment measurements. An
important measurement is the rate of alarms
1,500
annunciated to a single operator.
Figure 2 shows an overloaded alarm sys-
1,000
tem. The difference between the two lines is
the effect of including or removing only 10
individual high-rate nuisance alarms. This is 500
a common problem that is discussed later in
the article. 0
Acceptable range: 150 to 300 58 days
To respond to an alarm, an operator must
detect the alarm, investigate the conditions with the proper determination of the cause Figure 2. Removing a small
causing the alarm, decide on an action, take or causes. The diagnosis time varies based number of high-rate alarms
can have a large effect on the
the action and finally, monitor the process upon the operator’s experience and involve- alarm system’s overall profile
to ensure that the action taken resolves the ment in previous similar situations.
alarmed condition. These steps take time Process control graphics (human-ma-
and some must necessarily be executed se- chine interfaces; HMIs) play a major role in
quentially. Others can be performed in paral- effective detection of abnormal situations
lel as part of a response to several alarms and responses to them. Using effective
occurring simultaneously. HMIs, an operator can quickly and prop-
Given these steps, handling one alarm in erly ascertain the cause and corrective ac-
10 minutes (that is, approximately 150 over a tion for an abnormal situation. However, the
24-h period) can generally be accomplished quality of the HMI varies widely throughout
without the significant sacrifice of other op- the industry. Most HMI implementations
erational duties, and is considered likely to be are little more than a collection of numbers
acceptable. A rate greater than 150 per day sprinkled on a screen while showing a pip-
begins to become problematic. Up to two ing and instrumentation diagram (P&ID),
alarms per 10-minute period (~300 alarms/ making diagnosis much more difficult. For
day) are termed the “maximum manageable.” more discussion on this topic, search the
More than that may be unmanageable. Internet for the term “High-Performance
The acceptable alarm rates for small peri- HMI,” or see the comprehensive white
ods of time (such as 10 minutes or one hour) paper cited in Refs. 1 and 2.
depend on the specific nature of the alarm, As a result, the diagnosis and response
rather than the raw count. The nature of the to a simple high-tank-level alarm becomes
response varies greatly in terms of the de- quite complicated. Given the tasks involved,
mand upon the operator’s time. The duration it might only be possible to handle a few
of time required for an operator to handle an such alarms in an hour.
alarm depends upon the particular alarm. Other alarms are simpler, such as, “Pump
As an example, consider a simple tank with 412 should be running but has stopped.”
three inputs and three outputs. The tank’s The needed action is very direct: “Restart the
high-level alarm occurs. Consider all of the pump, or if it won’t restart, start the spare.”
possible factors causing the alarm and what Operators can handle several such alarms
the operator has to determine: as these in 10 minutes. It takes less time to
• Too much flow on inlet stream A, or B or C assess and work through the situation.
• Too much combined flow on streams Response to alarm rates of 10 alarms per
A-B, A-C, B-C or A-B-C 10 minutes (the threshold of a “flood”) can
• Not enough flow on outlet stream D, E or F possibly be achieved for short periods of time
• Not enough combined flow on streams — but only if the alarms are simple ones. And
D-E, D-F, E-F or D-E-F this does not mean such a rate can be sus-
• Several more additional combinations of tained for many 10-minute periods in a row.
the above inlet and outlet possibilities. During flood periods (Figure 3), operators are
The above situation takes quite a while likely to miss important alarms. Alarm rates
to diagnose, and involves observing trends per 10 minutes into the hundreds or more,
of all of these flows and comparing them to lasting for hours, are common. What are the
the proper numbers for the current process odds that the operator will detect the most
situation. The correct action varies highly important alarms in such a flood? Alarm
100
20
0
8 weeks
0
8 weeks
ing), fleeting (occurring and clearing in very
floods can make a difficult process situation short intervals), stale, duplicate and so forth.
much worse, and are often the precursors to Alarms with such behaviors are called “bad
major upsets or accidents. actors.” The most common cause of high
alarm rates is the misconfiguration of spe-
Averages can be misleading cific alarms, resulting in unnecessarily high
Alarm performance should generally be alarm occurrence rates. Commonly, 60–80%
viewed graphically rather than as a set of of the total alarm occurrences on a system
averages. Imagine that during one week, come from only 10–30 specific alarms. Chat-
your alarm system averaged 138 alarms per tering alarms and fleeting alarms are both
day and an average 10-minute alarm rate of common. Simply ranking the frequency of
0.96. That would seem to be well within the alarms will identify the culprits. Finding and
bounds of acceptability. But the data pro- correcting these rate-related nuisance be-
ducing those average numbers could look haviors will significantly reduce alarm rates
like that shown in Figure 4. with minimal effort.
The first flood lasted 40 minutes with 118 In the example data shown in Figure 6,
alarms. The second flood lasted 30 minutes 76% of all alarm occurrences came from
with 134 alarms. How many of those alarms only 10 individual configured alarms. In fact,
were likely to be missed? A simplistic answer the top two alarms make up 50% of the
(but good enough for this illustrative purpose) total load, with about 48,000 instances in
is to count the alarms that exceed 10 within 30 days. Alarms are never intentionally de-
any 10-minute period for the duration of each signed to annunciate so frequently, but they
flood, which, for the current example, would do. In this configuration, they would not per-
be a total of 182. In other words, despite form a useful function; rather, they would be
these seemingly great averages (many plant annoying distractions.
managers would consider these averages to Many of these were chattering alarms. In
be strong alarm-system performance and that summarizing 15 alarm-improvement projects
they would be happy to achieve), the alarm at power plants, the author’s employer found
pattern still puts the operators in the position that 52% of all alarm occurrences were as-
of likely missing almost 200 alarms. Missing sociated with chattering alarms. Proper ap-
so many alarms can result in improper opera- plication of alarm deadband and alarm on-
tor actions and undesirable consequences delay/off-delay time settings usually corrects
— perhaps quite significant ones. the chattering behavior. The calculations for
It is easy to plot such data, as in Figure 5. determining those settings are straightfor-
During an eight-week period, almost 21,000 ward (but beyond the scope of this article).
alarms were likely to be missed. A weekly Much more detailed information for solving
view of such data in this way will likely gain the all types of nuisance alarm problems can be
attention of management, whereas viewing found in Ref. 3.
the overall averages alone would indicate that
things are satisfactory when they are not. Alarm rationalization
The other cause of high alarm rates requires
Bad actor alarm reduction more effort to address. Most alarm systems
Many types of nuisance alarm behaviors are initially configured without the benefit of
exist, including chattering (rapidly repeat- a comprehensive “alarm philosophy” docu-
58 Chemical Engineering www.chemengonline.com march 2016
ment. This document sets out the rules Annunciated alarms per 10 minutes Figure 4. Different alarm data can gen-
70 erate similar average alarm rates, and the
for determining what kinds of situa- average rate may not tell the full story
tions qualify for alarm implementation. 60
Flood of
118 alarms Flood of
It specifies methods for consistently de- over 40 134 alarms
min. over 30
termining alarm priority, controlling alarm 50
min.
suppression, ongoing performance 40
analysis, management of change, and
dozens of other essential alarm-related 30
topics.
20
Systems created without such a doc-
ument are usually inconsistent collec- 10
tions of both “true alarms,” along with
many other items, such as normal sta- 0
7 days
tus notifications that should not use the
alarm system. Such non-alarms dimin- neously as an initial improvement effort
ish the overall effectiveness of the sys- with fast, high-impact results.
tem and diminish the operator’s trust in Step 1: Develop, adopt and maintain
it. They must be purged. While it may an alarm philosophy. A comprehensive
be easy to spot things that clearly have guideline for the development, imple-
no justification for being alarms by look- mentation and modification of alarms,
ing at the list of most frequent alarms, an alarm philosophy establishes basic
a comprehensive alarm rationalization is principles for a properly functioning
needed to ensure the consistency of the alarm system. It provides an optimum
overall alarm system. basis for alarm selection, priority set-
With alarm rationalization, every exist- ting, configuration, response, handling
ing alarm is compared to the principles methods, system monitoring and many
in the alarm philosophy document and other topics.
is either kept, modified or deleted. Set- Step 2: Collect data and benchmark
points or logical conditions are verified. the alarm system. Measuring the exist-
Priority is assigned consistently. New ing system against known, best-practice
alarms will be added, but the usual performance indicators identifies spe-
outcome of rationalization is a reduc- cific deficiencies, such as various types
tion in configured alarms by 50–75%. of nuisance alarms, uncontrolled sup-
Since the alarm-management prob- pression, and management-of-change
lem was identified in the early 1990s, issues. A baseline is established for im-
thousands of alarm systems have un- provements measurement.
dergone this process and achieved the Step 3: Perform “bad actor” alarm
desired performance. resolution. Addressing a few specific
After the bad actor reduction and the alarms can substantially improve an
rationalization steps, alarm rates are alarm system. Bad actor alarms, which
usually within the target limits. A typical can render an alarm system ineffective,
result is shown in Figure 7. Significant are identified and corrected to be con-
process upsets, particularly equipment sistent with the alarm philosophy. An
trips, may still produce some alarm ongoing program to identify and resolve
floods, which can be addressed in Step nuisance alarms is necessary.
6 listed below. Step 4: Perform alarm rationalization.
The 2009 publication of the ISA-18.2 Alarm rationalization is a comprehensive
Alarm Management Standard includes review of the alarm system to ensure it
both having an alarm philosophy docu- complies with the principles in the alarm
ment and performing alarm rationaliza- philosophy. This team-based effort re-
tion as mandatory items. For a compre- examines existing and potential alarms
hensive white paper on understanding configured on a system. Alarms to be
and applying ISA-18.2, see Ref. 4. added, deleted and reconfigured are
identified, prioritized and documented.
Alarm management work process The resulting alarm system has fewer
There is an efficient seven-step plan for configured alarms and is consistent and
improving an alarm system, proven in documented with meaningful priority
more than 1,000 improvement projects and setpoint values.
in plants throughout the world. Steps Step 5: Implement alarm audit and
1–3 are simple, and often done simulta- enforcement technology. Once an
3000
400
2000
200
1000
0 0
8 weeks
31 days
Figure 5. Despite sound averages for alarm rates, it can still Figure 7. Alarm rates can usually be brought into target lim-
be the case that many alarms could be missed during alarm its by alarm rationalization and bad-actor reduction steps
flood periods
alarm system is rationalized, its configura- program of system analyses that may in-
tion must not change without authoriza- clude KPI monitoring and the correction of
tion. Because DCS systems can be easily problems as they occur.
changed by a variety of sources, they often
require mechanisms that frequently audit Concluding remarks
(and enforce) the approved configuration. The various problems with alarm systems are
Step 6: Implement advanced alarm man- well recognized and there are proven solutions
agement. Certain advanced alarm capa- to these problems. The principles from these
bilities may be needed on some systems to solutions have been successfully applied to
address specific issues. For example, state- thousands of alarm systems worldwide. The
based alarming monitors the current process alarm management body of knowledge is ma-
state, and alarm settings are dynamically al- ture. Solving alarm-system problems simply
tered in predetermined ways to match the requires the will and effort to do so. n
alarming requirements of that process state. Edited by Scott Jenkins
Alarm flood suppression temporarily elimi-
nates the expected and distracting alarms References
from a unit trip, leaving the relevant alarms 1. Hollifield, B. and Perez, H. Maximize Operator Effectiveness: High
that assist the operator in managing that Performance HMI Principles and Best Practices, Part 1 of 2. PAS
Inc., Houston, 2015.
post-trip situation. Such advanced methods
2. Hollifield, B. and Perez, H. Maximize Operator Effectiveness: High
can ensure that the alarm system is effective Performance HMI Case Studies, Recommendations, and Standards,
even in abnormal situations. Part 2 of 2. PAS Inc., Houston 2015.
Step 7: Control and maintain the im- 3. Hollifield, B. and Habibi, E. The Alarm Management Handbook, 2nd
proved system. An effective alarm system Ed., PAS Inc., Houston 2010.
requires an ongoing and typically automated 4. Hollifield, B. Understanding and Applying the ANSI/ISA 18.2 Alarm
Management Standard. PAS Inc., Houston 2010.
Alarm
count Most frequent annunciated alarms Author
35,000 100 Bill Hollifield is the principal consultant at
30 days PAS Inc. (16055 Space Center Blvd., Suite
90
30,000 600, Houston, TX 77062; Phone: 281-286-
80
6565; Email: bhollifield@pas.com). He is re-
25,000 70 sponsible for alarm management and high-
Cumulative %
20,000
60 performance HMI. He is a member of the
50 ISA-18 Alarm Management committee, the
Ten alarms make
15,000 up 76% of the total ISA-101 HMI committee, and is a co-author of
40
alarm loading the Electric Power Research Institute’s (EPRI)
10,000 30 Alarm Management Guidelines. Hollifield is
20 also coauthor of the Alarm Management Handbook and The High
5,000
10
Performance HMI Handbook, along with many articles on these top-
ics. Hollifield has a dozen years of international, multi-company ex-
0 0 perience in all aspects of alarm management and effective HMI
Tag1.Alarm
Tag2Low1
Tag3.Alarm
Tag4.Alarm
Tag5.Low1
Tag6.Low1
Tag7.Alarm
Tag8.Low1
Tag9.High1
Tag10.Alarm
Wireless
Communication
in Hazardous Areas
Stephan Schultz
R. Stahl Consider these criteria in deciding where
wireless fits in today’s CPI plants and the
W
ireless communications have
great potential in the chemi-
cal process industries (CPI)
explosive atmospheres that permeate them
because they do away with
complex and costly cable installations In existing plants, power sources are application for wireless technology. As
and enable completely new applica- around nearly every corner, so the cost of yet, though, most reading devices
tions. And while a recent wave of suc- of wiring for power is not nearly as used for this purpose are handheld
cessful demonstrations has begun to significant as the cost of the wiring for terminals with a cable that curtails
emerge in the CPI (for more, see CE, the control signals themselves. their operation. Portable radio devices
Nov. 2009, p. 17–23), a number of hur- A look at typical routines in process capable of both acquiring data and
dles stand in the way of a completely plants will identify the potential ancil- passing it on via wireless link to MES
wireless Utopia. In most cases, the lary application areas with a view to (manufacturing execution system)
totally reliable, uncompromised avail- how and how much they may benefit. and ERP (enterprise resource plan-
ability of a production plant remains a Once a case is made for wireless tech- ning) servers save time and costs, and
paramount objective, and it will there- nology in general for these purposes, increase data reliability due to exact
fore likely take some more time before users are faced with various solutions to and nearly instant data acquisition.
radio transmissions of critical signals choose from for actual implementations. RFID tags can be expected to increase
in control loops take root. And last but not least, there are addi- their foothold in the CPI due to reli-
One impediment often cited as a limit tional safety considerations for applica- ability and safety benefits, since one
for wireless solutions is power. In fact, tions in hazardous areas. All of these key RFID advantage over barcodes is
many process applications basically aspects will be discussed in order to en- that even smudged and stained labels
rule out wireless field devices with- able users to make informed choices, or are still legible. Also, there are other
out an independent, onboard source to at least prime themselves for further convenient features that previous so-
of power. Granted, there have been a consultations with specialist manufac- lutions could not provide; for instance,
number of promising approaches in turers or systems solution providers. data can be written to the tags more
this regard, which are based on con- than once and it is possible to acquire
sumption-optimized electronic circuits application Areas several tags at the same time.
and alternative sources of power using Logistics and supply chain
accumulators or solar cells, or on so- State-of-the-art logistics solutions Maintenance and monitoring
called energy harvesting, where energy depend on systems that acquire data Anyone in the field who is servicing
is recovered from vibration, tempera- on the flows of goods with the highest a plant is likely to benefit from using
ture fluctuations, and so on. possible degree of precision, and pref- portable devices with a connection to
At the same time, there are a range erably at the very instant when stock a central management system, since
of ancillary functions in almost any items are taken out or replenished. doing so enables optimization of typi-
plant today for which wireless com- In the CPI, many raw materials and cal routines and measurements. For
munications truly are already a boon. products are transported in containers example, maintenance instructions
In these cases, power is not an insur- such as drums, tanks, intermediated can be automatically dispatched be-
mountable hurdle because the power bulk containers (IBCs), and so on. Most cause all relevant information can be
requirements are low enough to main- containers are marked with either provided via radio to a portable hand-
tain battery life of five or more years. barcodes or RFID (radio-frequency held device that service engineers can
Meanwhile, the use of wired power identification) tags. Acquiring RFID carry with them in the field. Staff are
should not be ruled out automatically. tag information is an obvious model then able to inspect equipment as
Chemical Engineering www.che.com may 2010 39
Feature Report
The main risk lies in the induction of less relevant. IEEE studies on elec-
electrical currents in metallic objects tromagnetic radiation in hazardous
or electronic circuits that are inade- areas have shown that even RF with
quately protected from electromagnetic power of 6 W can become a potential
interference (EMI). These currents hazard in terms of induction in metal
can result in excessively high tem- objects. Because of this danger, the IEC
peratures and the formation of sparks. 60079-0 2008 and the upcoming EN
Other dangers, such as direct ignition 60079-0 for continuous high frequency Figure 3. An external antenna in wire-
less units, such as this access point, is cur-
of an explosive atmosphere, are much sources limit the maximum permitted rently required to attain an individual ATEX
certification for use in hazardous areas
will actually work well, since the sig- • Customized solutions that
nal loss is otherwise excessive. meet your specific needs.
Because our systems are
Another possible option is the use of not “off the shelf,” they
external antennas. However, hazard- are always on the mark.
ous area requirements demand that • Reliable equipment that’s
special explosion-protected anten- proudly made in America.
nas have to be installed in this case. • Our Airtight Performance Guarantee™. We stand
They usually have to be designed for behind every system we engineer. And we say it
increased safety (Ex e) protection, be- in writing.
cause, in the event of a short circuit For more information about our custom-engineered
between the power supply and the pneumatic systems and solutions, call:
output or input stage in the RF device,
no excessively high currents or volt- 1-866-239-6936
ages are allowed to coincide with the or visit online at:
L
eaks in a chemical process in- “alphabet soup” of acronyms. The NEIC = National Enforcement
dustries (CPI) facility can run box on the right lists, for easy refer- Investigations Center
the gamut from creating a ence, the titles and acronyms that
NESHAP = National Emission
costly waste to prefacing a cat- will be used in this discussion.
Standard for Hazardous Air
astrophic failure.. They can be an Pollutants
annoyance, by creating pools of liq- Leak mechanisms
uid on concrete that can become a Eliminating the potential for leaks NSPS = New Source Performance
Standards
possible slipping hazard and house- is an integral part of the design
keeping problem, or a leak that can process that takes place at the very RCRA = Resource Conservation and
emit toxic vapors, causing various onset of facility design. It is woven Recovery Act
degrees of harm to personnel. In into the basic precept of the piping SOCMI = Synthetic organic chemical
some cases a leak may be a simple codes because it is such an elemen- manufacturing industry
housekeeping issue that goes into tal and essential component in the TSDF = Treatment, storage and
the books as a footnote indicating process of designing a safe and de- disposal facilities
that a repair should be made when pendable piping system. UST = Underground storage tank
resources are available. In other Piping systems, as referred to
cases it can become a violation of here, include pipe, valves and other VOC = Volatile organic
compounds
regulatory compliance with statu- inline components, as well as the
tory consequences, not to mention equipment needed to hold, move and
a risk to personnel safety and the process chemicals. Why then, if we (CA) is used as an applied factor
possible loss of capital assets. comply with codes and standards, in calculating, among other things,
Understanding the mechanisms and adhere to recommended indus- wall thickness in pipe and pressure
by which leaks can occur and priori- try practices, do we have to concern vessels. The CA value assigned to
tizing piping systems to be checked ourselves with leaks? Quite point- a material is theoretical and predi-
at specific intervals based on a few edly it is because much of what we cated on four essential variables:
simple factors is not only a prag- do in design is theoretical, such as material compatibility with the
matic approach to the preventive material selection for compatibility, fluid, containment pressure, tem-
maintenance of piping systems, but and because in reality, in-process perature of the fluid and velocity
is part of a CPI’s regulatory com- conditions and circumstances do of the fluid. What the determina-
pliance. This includes compliance not always perform as expected. tion of a CA provides, given those
under both the U.S. Environmen- Whether due to human error or variables, is a reasonable guess at
tal Protection Agency (EPA) Clean mechanical deficiencies, leaks are a uniform rate of corrosion. And
Air Act (CAA; 40CFR Parts 50 to a mechanism by which a contained given that, an anticipated loss of
52) and the Resource Conservation fluid finds a point of least resistance material can be assumed over the
and Recovery Act (RCRA; 40CFR and, given time and circumstances, theoretical lifecycle of a pipeline
Parts 260 to 299). We will get into breaches its containment. What we or vessel. It allows a reasonable
more detail with these regulations, look into, somewhat briefly, are two amount of material to be added into
as well as the leak detection and general means by which leaks can the equation, along with mechani-
repair (LDAR) requirement within occur; namely corrosion and me- cal allowances and a mill tolerance
the above mentioned regulations, as chanical joint deficiencies. in performing wall thickness cal-
we move through this discussion. Corrosion. Corrosion allowance culations. The problem is that be-
44 Chemical Engineering www.che.com May 2014
Table 1. Elements of a Model LDAR Program Pump, compressor and agitator
seals can develop leaks where shaft
Written LDAR compliance First attempt at repair
misalignment plays a part. If the
Training Delay of repair compliance assurance shaft is not installed within recom-
LDAR audits Electronic monitoring and storage of data mended tolerances or if it becomes
misaligned over time there is a
Contractor accountability QA/QC of LDAR data good possibility the seal will begin
Internal leak definitions Calibration/calibration drift assessment to fail.
Less frequent monitoring Records maintenance The LDAR program
Promulgated in 1970 and amended
yond the design, engineering, and areas for abnormal loss of wall in 1977 and 1990, the Clean Air
construction phase of building a thickness, hydrogen stress-corro- Act requires that manufactur-
facility, the in-service reality of cor- sion cracking (HSCC), and others. ers producing or handling VOCs
rosion can be very different. The LDAR program does not develop and maintain an LDAR
Corrosion, in the majority of specify the need to check anything program in accordance with the
cases, does not occur in a uniform other than mechanical joints for po- requirements set forth under the
manner. It will most frequently tential leaks. Monitoring pipe and Clean Air Act. This program moni-
occur in localized areas in the form vessel walls, particularly at welds tors and documents leaks of VOCs
of pits, as erosion at high-impinge- that come in contact with corrosive in accordance with Method 21 —
ment areas, as corrosion under chemicals, is a safety consideration Determination of Volatile Organic
insulation, at heat-affected zones and practical economics. Perform- Compound Leaks.
(HAZ) where welding was improp- ing cursory examinations for such Table 1 provides a listing of key
erly performed, causing a localized points of corrosion where the po- elements that should be contained
change to the mechanical or chemi- tential exists should be made part in an LDAR program. Those ele-
cal properties of the material, and of any quality assurance or quality ments are described as follows:
in many other instances in which control (QA/QC) and preventive Written LDAR compliance. Com-
unforeseen circumstances create maintenance program. pile a written procedure declaring
the potential for corrosion and the Mechanical joints and open- and defining regulatory require-
opportunity for leaks in the pipe ended pipe. Mechanical joints ments that pertain to your specific
itself or in a vessel wall. Because can include such joining methods facility. This should include record-
of that incongruity, corrosion is an as flanges, unions, threaded joints, keeping certifications; monitoring
anomaly that, in reality, cannot valve bonnets, stem seals and clamp and repair procedures; name, title,
wholly be predicted. assemblies. It can also include and work description of each person-
Corrosion-rate values found in pump, compressor and agitator nel assignment on the LDAR team;
various published resources on the seals. Other potential points of tran- required procedures for compiling
topic of material compatibility are sient emissions include open-ended test data; and a listing of all process
based on static testing in which a piping, such as drains, vents, and units subject to federal, state and
material coupon is typically set in the discharge pipe from a pressure- local LDAR regulations.
a vile containing a corrosive chemi- relief device. Any of these joints or Training. Assigned members of
cal. This can be done at varying interfaces can be considered poten- the LDAR team should have some
temperatures and in varying con- tial leak points and require both experience base that includes work
centrations. After a period of time, monitoring and record-keeping doc- performed in or around the types of
the coupon is pulled and the rate umentation in compliance with the piping systems they will be testing
of corrosion is assessed. That is a EPA’s LDAR program. and monitoring under the LDAR
simplification of the process, but Mechanical joints can leak due to program. Their training should in-
you get the point. When a material improper assembly, insufficient or clude familiarization with Method
of construction (MOC) and a po- unequal load on all bolts, improp- 21 and also training as to the cor-
tentially corrosive chemical come erly selected gasket type, sufficient rect procedure for how to examine
together in operational conditions, pressure or temperature swings the various interface connections
the theoretical foundation upon that can cause bolts to exceed their they will be testing. They should
which the material selection was elastic range (diminishing their also receive training on the test
based becomes an ongoing realtime compressive load on the joint), and instrument they will be using and
assessment. This means that due an improperly performed “hot-bolt- how to enter the test data in the
diligence needs to be paid to exam- ing” procedure in which in-service proper manner. All of this needs to
ining areas of particular concern, bolts are replaced while the pipeline be described in the procedure.
depending on operating conditions, remains in service. “Hot bolting” is LDAR audits. An internal audit
such as circumferential pipe welds not a recommended procedure, but team should be established to en-
for cracking, high-impingement is nonetheless done on occasion. sure that the program is being car-
Chemical Engineering www.che.com May 2014 45
Cover Story
ried out on a routine basis in an ef- ity has consistently demonstrated monitoring equipment should be
ficient and comprehensive manner good performance under monthly made at the end of each monitor-
in accordance with the written pro- testing, then the frequency of test- ing work shift using approximately
cedures. A third-party audit team is ing could be adjusted to a quarterly 500 ppm of calibration gas. If, after
brought in every few years to con- test frequency. the initial calibration, drift assess-
firm that internal audits are being First attempt at repair. Upon de- ment shows a negative drift of more
carried out in the proper manner tection of a leak, most rules will re- than 10% from the previous cali-
and that all equipment that should quire that a first attempt be made bration, all components that were
be included in the monitoring is to repair the leak within five days tested since the last calibration
listed as such. It also ensures that of detection; if unsuccessful, any fol- with a reading greater than 100
the tests are being carried out prop- low-up attempts need to be finalized ppm should be re-tested. Re-test all
erly and that the test results are within 15 days. Should the repair pumps that were tested since the
entered properly. remain unsuccessful within the 15- last calibration having a reading of
Contractor accountability. day time period, the leak must be greater than 500 ppm.
When selecting an outside con- placed on a “delay of repair” list and Records maintenance. Internal
tractor to perform internal LDAR a notation must be made for repair electronic record-keeping and re-
audits for a facility or when bring- or component replacement during porting is an essential component to
ing in an outside contractor to in- the next shutdown of which the a well-implemented LDAR program.
spect the work of the internal audit leaking component is a part. It is an indication to the NEIC that
team, it is recommended that the Delay of repair compliance as- every effort is being made to comply
contract be written in a manner surance. Placing a repair item on with the regulations pertinent to a
that places appropriate responsi- the “delay of repair” list gives assur- facility. It provides ready access to
bility on that contractor. In doing ances that the item justifiably be- the personnel associated with the
so there should be penalties de- longs on the list, that a plan exists program, the test data, leak repair
scribed and assessed as a result to repair the item, and that parts reports and so on.
of insufficient performance or in- are on hand to rectify the problem.
accurate documentation of pre- It is suggested that any item being Testing for leaks
scribed testing and documentation listed in the “delay of repair” list au- Results, when using a leak detec-
procedures. Expectations should tomatically generate a work order tion monitor, are only as accurate
be well defined and any deviation to perform the repair. as its calibration and the manner in
from those prescribed norms by a Electronic monitoring and stor- which it is used. Calibration is dis-
third-party contractor should con- age of data. Entering leak-test cussed in the next section, “Method
stitute a breach of contract. In all data into an electronic database 21.” To use the monitor correctly, the
fairness, both parties must under- system will help in retrieving such auditor will need to place the nozzle
stand exactley what those expecta- data and in utilizing them in ways or end of the probe as close as pos-
tions are. that help provide reports highlight- sible to the flange, threaded joint, or
Internal leak definitions. Inter- ing areas of greater concern to areas seal interface as follows:
nal leak definitions are the maxi- of lesser concern. Such information • In the case of a flange joint test:
mum parts per million, by volume can help direct attention and re- 180 deg around perimeter of the
(ppmv) limits acceptable for valves, sources away from areas of least flange joint at their interface
connectors and seals, as defined by concern, while mobilizing resources • In the case of a threaded joint test:
the CAA regulation governing a fa- to areas of greater concern. This en- 180 deg around perimeter of inter-
cility. For example, a facility may be ables a much more efficient use of face of the male/female fit-up
required to set an internal leak-def- information and resources. • If it is a coupling threaded at both
inition limit of 500 ppm for valves QA/QC of LDAR data. A well ends, check both ends 180 deg
and connectors in light liquid or gas/ written LDAR program will include around the perimeter
vapor fluid service and 2,000 ppm a QA/QC procedure defining the • If it is a threaded union, then
internal leak definition for pumps process by which it is assured that check both ends and the body nut
in light liquid or gas/vapor fluid Method 21 is being adhered to, and 180 deg around the perimeter
service. “Light liquid” is defined that testing is being carried out in • In the case of a valve test:
as a fluid whose vapor pressure is the proper manner and includes the 180 deg around perimeter of
greater than 0.044 psia at 68°F. proper equipment and components. all end connections if anything
Less frequent monitoring. Under This also includes the maintenance other than welded
some regulations it is allowed that of proper documentation. 180 deg around perimeter of
a longer period between testing is Calibration/calibration-drift body flange
acceptable if a facility has consis- assessment. LDAR monitoring 180 deg around perimeter of
tently demonstrated good perfor- equipment should be calibrated in body/bonnet interface
mance (as defined in the applicable accordance with Method 21. Cali- 180 deg around perimeter of
regulation). For example, if a facil- bration-drift assessment of LDAR stem packing at stem
46 Chemical Engineering www.che.com May 2014
National Cleanup Backlog
(confirmed releases, cleanups completed) 160,000
140,000
136,265
102,798
191,242
113,919
108,766
100,165
93,123
87,983
40,000 ization detector (FID) theoretically
measures the total carbon content
20,000 of the organic vapor sampled. The
photoionization detector (PID)
0
2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 uses ultraviolet light to ionize the
organic vapors. With both detec-
Fiscal year
tors, the response will vary with
Figure 1. Progress is slowly being made to clean up leaking underground storage the functional group in the organic
tanks under the RCRA program compounds. PIDs have been used to
detect equipment leaks in process
• In the case of a rotating equipment calibration. One is referred to as a units in SOCMI facilities, particu-
shaft seal test: 180 deg around the “zero gas,” defined as air with less larly for compounds such as form-
perimeter of the interface of the than 10 ppmv (parts per million aldehyde, aldehydes and other oxy-
seal and the shaft by volume) VOC. The other cali- genated chemicals that typically do
bration gas, referred to as a “refer- not provide a satisfactory response
Method 21 ence gas,” uses a specified reference on a FID-type unit.
Method 21, under 40 CFR Part 60, compound in an air mixture. The Operation of the non-dispersive
Appendix A, provides rules with concentration of the reference com- infrared (NDIR) detector is based
respect to how VOCs are moni- pound must approximately equal on the principle that light absorp-
tored and measured at potential the leak definition specified in the tion characteristics vary depending
leak points in a facility. Those po- regulation. The leak definition, as on the type of gas. Because of this,
tential leak points include, but are mentioned above, is the threshold NDIR detection can be subject to
not limited to: valves, flanges and standard pertinent to the govern- interference due in large measure
other connections; pumps and com- ing regulation. to such constituents as water vapor
pressors; pressure-relief devices; and CO2, which may absorb light
process drains; open-ended valves; Monitoring devices at the same wavelength as the tar-
pump and compressor seals; de- A portable VOC-monitoring device geted compound. This type of detec-
gassing vents; accumulator vessel will typically be equipped with a tor is typically confined to the de-
vents; agitator seals and access door rigid or flexible probe. The end of tection and measurement of single
seals. It also describes the required probe is placed at the leak inter- components. Because of that pro-
calibration process in setting up the face of a joint, such as a flange, clivity, good or bad, the wavelength
monitoring device. Essentially any threaded connection or coupling, at which a certain targeted com-
monitoring device may be used as or at the interface of a pump, com- pound absorbs infrared radiation,
long as it meets the requirements pressor, or agitator seal where it having a predetermined value, is
set forth in Method 21. interfaces with the shaft. With its preset for that specific wavelength
Cylinder gases used for calibrat- integral pump, the device, when through the use of optical filters. As
ing a monitoring device need to be switched on, will draw in a contin- an example, if the instrument was
certified to be within an accuracy uous sample of gas from the leak- set to a wavelength of 3.4 microm-
of 2% of their stated mixtures. It is interface area into the monitoring eters, the device could detect and
recommended that any certification device. The instrument’s response measure petroleum fractions, such
of this type be filed in either digital or screening value is a relative as gasoline and naphtha.
form or at the very least as a hard measure of the sample’s concentra- The combustion-type analyzer is
copy. There should also be a speci- tion level. The screening value is designed to measure either thermal
fied shelf life of the contents of the detected and displayed in parts per conductivity of a gas or the heat pro-
cylinder. If the shelf life is exceeded, million by volume, or if the instru- duced as a result of combustion of the
the contents must be either re-ana- ment is capable and the degree of gas. Referred to as hot-wire detectors
lyzed or replaced. accuracy needed, in parts per bil- or catalytic oxidizers, combustion-
Method 21 goes on to define how lion by volume (ppbv). type monitors are nonspecific for
to test flanges and other joints, as The detection devices operate on gas mixtures. If a gas is not readily
well as pump and compressor seals a variety of detection principles. combustible, similar in composition
and various other joints and inter- The most common are ionization, to formaldehyde and carbon tetra-
faces with the potential for leaks. infrared absorption and combus- chloride, there may be a reduced re-
There are two gases required for tion. Ionization detectors operate sponse or no response at all.
Chemical Engineering www.che.com May 2014 47
Table 2 – Federal Regulations That Require a Formal LDAR
Cover Story Program With Method 21
40 CFR Regulation Title
Part Subpart
Due to the variability in the sen- 60 VV SOCMI VOC Equipment Leaks NSPS
sitivity of the different monitoring
60 DDD Volatile Organic Compound (VOC) Emissions from the Poly-
devices, the screening value does mer Manufacturing Industry
not necessarily indicate the actual
60 GGG Petroleum Refinery VOC Equipment Leaks NSPS
total concentration at the leak in-
terface of the compound(s) being 60 KKK Onshore Natural Gas Processing Plant VOC Equipment
Leaks NSPS
detected. The leak interface is the
immediate vicinity of the joint 61 J National Emission Standard for Equipment Leaks (Fugitive
being tested — the point at which Emission Sources) of Benzene
the end of the probe is placed. Re- 61 V Equipment Leaks NESHAP
sponse factors (RFs), determined 63 H Organic HAP Equipment Leak NESHAP (HON)
for each compound by testing or
63 I Organic HAP Equipment Leak NESHAP for Certain Processes
taken from reference sources, then
correlate the actual concentration 63 J Polyvinyl Chloride and Copolymers Production NESHAP
of a compound to that of the con- 63 R Gasoline Distribution Facilities (Bulk Gasoline Terminals and
centration detected by the moni- Pipeline Breakout Stations)
toring device. As mentioned previ- 63 CC Hazardous Air Pollutants from Petroleum Refineries
ously, the monitoring device must 63 DD Hazardous Air Pollutants from Off-Site Waste and Recovery
first be calibrated using a certified Operations
reference gas containing a known
63 SS Closed Vent Systems, Control Devices, Recovery Devices
compound at a known concentra- and Routing to a Fuel Gas System or a Process
tion, such as that of methane and
63 TT Equipment Leaks – Control Level 1
isobutylene. RFs at an actual con-
centration of 10,000 ppmv have 63 UU Equipment Leaks – Control Level 2
been published by the EPA in a 63 YY Hazardous Air Pollutants for Source Categories: Generic
document entitled “Response Fac- Maximum Achievable Control Technology Standards
tors of VOC Analyzers Calibrated 63 GGG Pharmaceuticals Production
with Methane for Selected Organic 63 III Hazardous Air Pollutants from Flexible Polyurethane Foam
Chemicals.” Production
Method 21 requires that any se-
63 MMM Hazardous Air Pollutants for Pesticide Active Ingredient
lected detector meet the following Production
specifications:
63 FFFF Hazardous Air Pollutants: Miscellaneous Organic Chemical
• The VOC detector should respond Manufacturing
to those organic compounds being
processed (determined by the RF) 63 GGGGG Hazardous Air Pollutants: Site Remediation
• Both the linear response range 63 HHHHH Hazardous Air Pollutants: Miscellaneous Coating Manufac-
and the measurable range of the turing
instrument for the VOC to be 65 F Consolidated Federal Air Rule — Equipment Leaks
measured and the calibration gas 264 BB Equipment Leaks for Hazardous Waste TSDFs
must encompass the leak defini-
265 BB Equipment Leaks for Interim Status Hazardous Waste TSDFs
tion concentration specified in the
regulation
• The scale of the analyzer meter Federal regulations program and help establish a com-
must be readable to ±2.5% of There are federal regulations that prehensive and detailed procedure.
the specified leak definition pertain to monitoring for VOCs
concentration and require the implementation of RCRA
• The analyzer must be equipped a formal LDAR program in concert The Solid Waste Disposal Act of
with an electrically driven pump with the rules of Method 21. There 1965 was amended in 1976 to in-
so that a continuous sample is are other federal regulations that clude the Resource Conservation
provided at a nominal flowrate of require the rules of Method 21, but and Recovery Act (RCRA), which
between 0.1 and 3.0 L/min do not require a formal LDAR pro- encompassed the management of
• The analyzer must be intrinsi- gram. Tables 2 and 3 list those vari- both hazardous waste and solid
cally safe for operation in explo- ous regulations. waste. Prompted further by an ever
sive atmospheres It is the manufacturer’s responsi- increasing concern of underground
• The analyzer must be equipped bility to make the proper determina- water contamination, this act was
with a probe or probe extension tion as to what regulations it needs again amended in 1984 to address
not to exceed 0.25 in. outside di- to comply with. Those specific regu- underground storage tanks (USTs)
ameter with a single end opening lations, coupled with the Method 21 and associated underground piping
for sampling requirements, will define the LDAR under Subtitle I. This Amendment
48 Chemical Engineering www.che.com May 2014
Table 3 – Federal Regulations that Require the Use of Method 21
But Not a Formal LDAR Program
40 CFR Regulation Title
Part Subpart
60 XX Bulk Gasoline Terminals • Oversee cleanups of petroleum re-
60 QQQ VOC Emissions from Petroleum Refinery Wastewater Systems leases by responsible parties
60 WWW Municipal Solid Waste Landfills • Enforce cleanups by recalcitrant
parties
61 F Vinyl Chloride
• Pay for cleanups at sites where
61 L Benzene from Coke By-Products the owner or operator is unknown,
61 BB Benzene Transfer unwilling, or unable to respond,
61 FF Benzene Waste Operations or those that require emergency
action
63 G Organic Hazardous Air Pollutants from SOCMI for Process
Vents, Storage Vessels, Transfer Operations, and Wastewater
• Conduct inspections and other re-
lease prevention activities
63 M Perchloroethylene Standards for Dry Cleaning In Figure 1 the progress being
63 S Hazardous Air Pollutants from the Pulp and Paper Industry made by the program can readily
63 Y Marine Unloading Operations be seen. In 2002, RCRA was looking
at 142,709 LUST sites — sites that
63 EE Magnetic Tape Manufacturing Operations
were flagged for cleanup. Through-
63 GG Aerospace Manufacturing and Rework Facilities out the following nine years, 2002
63 HH Hazardous Air Pollutants from Oil and Gas Production through 2011, 54,726 of those sites
Facilities were cleaned, leaving 87,983 still
63 OO Tanks — Level 1 targeted for cleanup.
63 PP Containers Within the RCRA program there
are requirements that impact de-
63 QQ Surface Impoundments
sign, fabrication, construction, loca-
63 VV Oil/Water, Organic/Water Separators tion, monitoring and operation of
63 HHH Hazardous Air Pollutants from Natural Gas Transmission and USTs and associated underground
Storage piping. The EPA has provided a
63 JJJ Hazardous Air Pollutant Emissions: Group IV Polymers and number of sites on the internet that
Resins provide a great deal of information
63 VVV Hazardous Air Pollutants: Publicly Owned Treatment Works on the various CFR Parts. 40 CFR
Part 260 contains all of the RCRA
65 G CFAR — Closed Vent Systems
regulations governing hazardous
264 AA Owners and Operators of Hazardous Waste Treatment, Stor- waste identification, classification,
age, and Disposal Facilities — Process Vents generation, management and dis-
264 CC Owners and Operators of Hazardous Waste Treatment, posal.
Storage and Disposal Facilities — Tanks, Surface Impound- Listed wastes are divided into the
ments, Containers
following group designations:
265 AA Interim Standards for Owners and Operators of Hazardous • The F group — non-specific source
Waste Treatment, Storage, and Disposal Facilities — Process wastes found under 40 CFR
Vents
261.31
265 CC Interim Standards for Owners and Operators of Hazardous • The K group — source-specific
Waste Treatment, Storage, and Disposal Facilities — Tanks,
Surface Impoundments, Containers wastes found under 40 CFR
261.32
270 B Hazardous Waste Permit Program — Permit Application
• The P and U group — discarded
270 J Hazardous Waste Permit Program — RCRA Standardized Per- commercial chemical products
mits for Storage Tanks and Treatment Units found under 40 CFR 261.33
Characteristic wastes, which exhibit
regulates the construction, moni- by current UST regulations to pre- one or more of four characteristics
toring, operating, reporting, record- vent and detect releases and were defined in 40 CFR Part 261 Subpart
keeping, and financial respon- performing the necessary UST sys- C are as follows:
sibility for USTs and associated tem operation and maintenance. • Ignitability, as described in 40
underground piping that handle In 1986, the Leaking Under- CFR 261.21
petroleum and hazardous fluids. ground Storage Tank (LUST) Trust • Corrosivity, as described in 40
As of 2011, there were 590,104 Fund was added to the RCRA pro- CFR 261.22
active tanks and 1,768,193 closed gram. The trust financing comes • Reactivity, as described in 40 CFR
tanks in existence in the U.S. Of the from a 0.1¢ tax on each gallon of 261.23
still active tanks, 70.9% were under motor fuel (gasoline, diesel or bio- • Toxicity, as described in 40 CFR
significant operational compliance. fuel blend) sold nationwide. The 261.24
This means that they were using LUST Trust Fund provides capital Table 4 provides a listing of ad-
the necessary equipment required to do the following: ditional CFR parts that further
Chemical Engineering www.che.com May 2014 49
Table 4 – Resource Conservation and Recovery Act (RCRA)
Feature
Cover Story
Report Information
40 CFR Part Regulation Title
260 Hazardous Waste Management System: General
define the regulations under the 261 Identification and Listing of Hazardous Waste
Resource Conservation and Recov-
ery Act. 262 Standards Applicable to Generators of Hazardous Waste
264 Standards for Owners and Operators of Hazardous Waste Treat-
Final remarks ment, Storage and Disposal Facilities
I am fervently against overregula-
265 Interim Status Standards for Owners and Operators of Hazardous
tion and watch with keen interest Waste Treatment, Storage and Disposal Facilities
the unfolding debate occurring on
Capitol Hill over the amendment 266 Standards for the Management of Specific Hazardous Wastes
and Specific Types of Hazardous Waste Management Facilities
to the Toxic Substances Control
Act (TSCA) for example. But the 267 Standards for Owners and Operators of Hazardous Waste Facili-
improved safety, clean air, clean ties Operating Under a Standardized Permit
water, and cost savings realized 270 EPA Administered Permit Programs: The Hazardous Waste Permit
from the CAA and RCRA programs Program
are four major returns on invest-
272 Approved State Hazardous Waste Management Programs
ment that come back to a manufac-
turer from the investment in a good 273 Standards for Universal Waste Management
leak-detection program. Whether
279 Standards for the Management of Used Oil
monitoring and repairing leaks
above ground, in accordance with 280 Technical Standards and Corrective Action Requirements for
the CAA, or below ground, in accor- Owners and Operators of Underground Storage Tanks (UST)
dance with the RCRA, it is, simply 281 Approval of State Underground Storage Tank Programs
put, just good business. As alluded
to at the outset of this article, leaks 282 Approved Underground Storage Tank Programs
RG LeaseFleet Ad_4.5625 x 4.875.pdf 1 4/7/14 11:47 AM
in hazardous-fluid-service pip-
ing systems have served, in many
cases, as an early-warning indicator
of something much worse to come.
At the very least, such leaks can
contribute to air pollution, ground-
water contamination, lost product
revenue, housekeeping costs, and a
risk to personnel — a few things we
can all live without. ■
Edited by Gerald Ondrey INTRODUCING THE WORLD’S LARGEST FLEET OF
SUCCESSFULLY TESTED BLAST-RESISTANT BUILDINGS.
Author
W. M. (Bill) Huitt has been No matter how many you need, how big you need them to be or when
involved in industrial pip-
ing design, engineering and you need them, get proven protection from the safety authority.
construction since 1965. Posi-
tions have included design en-
gineer, piping design instruc-
tor, project engineer, project
supervisor, piping depart-
ment supervisor, engineering
manager and president of W. • F O R M E R LY A B O X 4 U •
M. Huitt Co. (P.O. Box 31154,
St. Louis, MO 63131-0154;
Phone: 314-966-8919; Email: wmhuitt@aol.
com), a piping consulting firm founded in 1987.
His experience covers both the engineering and
construction fields and crosses industry lines
to include petroleum refining, chemical, petro-
chemical, pharmaceutical, pulp & paper, nuclear
power, biofuel and coal gasification. He has writ-
ten numerous specifications, guidelines, papers,
and magazine articles on the topic of pipe design
and engineering. Huitt is a member of the In-
ternational Society of Pharmaceutical Engineers
(ISPE), the Construction Specifications Institute
(CSI) and the American Society of Mechani-
cal Engineers (ASME). He is a member of the
B31.3 committee, a member of three ASME-BPE
subcommittees and several task groups, ASME
Board on Conformity Assessment for BPE Certi-
fication where he serves as vice chair, a member
of the American Petroleum Institute (API) Task
Group for RP-2611, serves on two corporate spec-
ification review boards, and was on the Advisory 855.REDGUARD
Board for ChemInnovations 2010 and 2011 a
multi-industry conference & exposition.
Circle 1 on p. 76 or go to adlinks.che.com/50976-01
Chemical Engineering www.che.com May 2014 51
Environmental Manager
F
ire is a primary and very real
threat to people, equipment
and facilities in the chemical
process industries (CPI), es-
pecially in the refining and storage of
petrochemicals. The consequences
of failing to detect flames, combusti-
ble gas leaks or flammable chemical
spills can have dire consequences,
including loss of life and catastrophic
plant damage.
The monitoring of flame hazards is
mandated by the U.S. Occupational
Safety and Health Administration
(OSHA; Washington, D.C.; www.
osha.gov) through its comprehen-
sive Process Safety Management
(PSM) federal regulation. Interna-
tionally, the European Union (E.U.)
splits gas and flame safety respon-
sibilities between E.U. directives and
European standards organizations, FIGURE 2. Flame detectors, such as those shown here, implement ultraviolet and infrared detection
including the European Committee technologies
for Electrotechnical Standardization
(Cenelec; Brussels, Belgium; www. to train employees to follow related including ultraviolet (UV) and infra-
cenelec.eu), the International Elec- safety procedures consistently. In red (IR) spectroscopy and visual
trotechnical Commission (IEC; Ge- either case, it is important to under- flame imaging. The source of flames
neva, Switzerland; www.iec.ch) and stand the many different sources of in CPI plants is typically fueled by
several other bodies. flame hazards, the detection sensor hydrocarbons, which when sup-
Many accidents are the result of ei- technologies that can warn of immi- plied with oxygen and an ignition
ther failing to implement these stan- nent danger and the proper location source, produce heat, carbon diox-
dards properly with suitable flame- of flame detectors in today’s com- ide and other products of combus-
detection equipment or the failure plex chemical plants. tion. Intense flames emit visible, UV,
In the petrochemical plant environ- and IR radiation (Figure 1). Flame
Ultraviolet Visible Infrared
ment, the range of potential flam- detectors are designed to detect
mable hazards is expansive and the emission of light at specific
growing as materials and processes wavelengths, allowing them to dis-
become more complex. These haz- criminate between flames and false
Relative energy
Sun's
energy ards have led to the development of alarm sources.
reaching
the more sophisticated combustible-gas
earth and flame-sensing technologies with Flame-sensing technologies
embedded intelligence that can bet- The flame safety industry has de-
ter detect the most common indus- veloped four primary optical flame-
300 nm 400 nm 800 nm 4-5 m
trial fire sources, some of which are sensing technologies: UV, UV/IR,
Wavelength listed in Table 1. multi-spectrum infrared (MSIR),
and visual flame imaging (Figure 2).
FIGURE 1. Flame detectors can detect light emis-
sions at specific wavelengths across the UV, vis-
Principles of flame detection These sensing technologies are all
ible and IR spectrum to distinguish between actual Industrial process flame detectors based on line-of-sight detection of
flames and false alarm sources detect flames by optical methods, radiation emitted by flames in the
70 Chemical Engineering www.chemengonline.com May 2016
Table 1. Common Industrial Input layer Hidden layer Output layer
Fire Sources
Alcohols Liquefied natural gas (LNG)
Diesel fuels Liquefied petroleum gas (LPG) Sensor 1
Gasoline Paper
Kerosene Textiles
Jet fuels Solvents Sensor 2
Output
Ethylene Sulfur
Hydrogen Wood
Sensor 3
Circle 07 on p. 94 or go to adlinks.chemengonline.com/61495-07
Integrated
Risk-Management Matrices
An overview of the tools available to reliability professionals for making their
organization the best-in-class
In Brief
Reliability,
historically
Reliability, today
Risk-mitigation
approaches
How do we measure
risk?
S
organization combines state-of-the-art theory,
ince the 1960s, process facility op- software and condition-monitoring techniques
erators have made concerted ef- with a strong collaboration of departments
forts to improve the overall reliability and associated personnel. An independent
and availability of their plants. From risk-based inspection (RBI) program or reli-
reliability theory to practical advancements ability-centered maintenance (RCM) program
in non-destructive examination and condi- no longer suffices as cutting-edge. Rather, the
tion-monitoring techniques, the industry has inspection department (power users of RBI)
significantly evolved and left key operations and maintenance department (power users
personnel with more tools at their disposal of RCM) are integrating with process, opera-
than ever before. However, this deeper ar- tions, capital projects and other teams to form
senal of tools, coupled with more stringent an overall reliability work process for the suc-
regulatory scrutiny and internal business cess of the plant.
pressure, introduces a heightened expecta- To highlight reliability’s growing prominence
tion of performance. Now, more than ever, within process facilities, this article addresses
companies recognize that best-in-class reli- the following:
ability programs not only save lives but in- • A brief history of reliability practices in the
crease the bottom line. These programs are 20th and 21st centuries
also one of the foremost “levers” for C-level • Examples of current reliability program tools
personnel to pull when trying to contend in a • A characterization of three different
risk-mitigation applications that cycle. As a result, several different de- Reliability today
are currently applied in process partments and individuals cooperated With the greater expectation on to-
facilities to ensure they attained reliability. day’s programs, department manag-
• The case for ensuring these risk The concept of RCM pushed ers (including reliability, mechanical-
mitigation frameworks are working through some industries quicker integrity or maintenance managers)
together than others. While it started with the face a powerful, but often intimidat-
• The value of key performance airlines, it flowed quickly into power ing array of tools available to them for
indicators (KPIs) in providing generation, petrochemical and petro- improving their reliability programs.
transparency and accountability leum-refining operations thereafter. Examples are listed in Table 1.
to the effectiveness of these risk Fast-forward to 1992, and another While this only represents a sub-
mitigation frameworks facet, called process-safety manage- set of the options available to the
ment (PSM), was introduced into the re- manager, all of these activities aim at
Reliability, historically liability picture. In response to a growing doing the following:
When one thinks about process reli- perception of risk related to hazardous 1. Reducing the risk of unplanned
ability, a variety of definitions come processes, the Occupational Safety downtime.
to mind. However, it has come a and Health Administration (OSHA) is- 2. Limiting safety and environmental
long way since the early 20th cen- sued the Process Safety Standard, risk.
tury. From the 1920s to the 1950s, OSHA 1910.119, which includes the 3. Ensuring compliance with regula-
reliability went from being classified following 14 required elements: tory standards.
as “repeatability” (how many times • Process-safety information 4. Doing steps one through three for
could the same results repeat) to de- • Process hazard analysis the least cost possible.
pendability (hours of flight time for an • Operating procedures To summarize, the goal of these
engine), to a specific, repeatable re- • Training managers is to put a plan in place
sult expected for a duration of time. • Contractors and execute a plan that identifies and
Through the 1950’s age of industri- • Mechanical integrity mitigates risks as efficiently as pos-
alization, reliability’s evolving definition • Hot work sible. To do that, one has to system-
was still very much focused on de- • Management of change atically identify those risks in addition
sign and not as much on operations • Incident investigation to the level to which those risks must
or maintenance. Then in the 1960s, • Compliance audits be mitigated. If this is done correctly,
the airline industry introduced the • Trade secrets the design, inspections, preventative
concept of reliability centered mainte- • Employee participation maintenance, operational strategies,
nance (RCM), pushing the idea that • Pre-startup safety review and other program facets should
the overall reliability of a system in- • Emergency planning & response all be aligned in attaining steps one
cluded not only the design, but also The intent of the regulation was to through four.
the operations and maintenance of limit the overall risk related to danger-
that system. In other words, reliability ous processes, and “raise the bar” Risk-mitigation approaches
engineering was now stretching into for compliance expectation for facili- Since the 1960s, there have been
other departments, mandating that ties with these “covered” processes. substantial efforts on figuring out
the overall risk of failure was tied to At that point, it became law to fulfill how to best characterize both
multiple aspects of the asset’s life- these 14 elements, and to ignore downtime and loss-of-containment
66 Chemical Engineering www.chemengonline.com may 2016
team on the identification of process
PHA/HAZOP/QRA deviations that can lead to undesir-
(quantitative risk assessment)
able consequences, the risk ranking
of those deviations, and the assign-
ment of actions to either lower the
Consequence of failure
RBI Extreme
probability of those failures or the
High consequence if the failures do occur.
While a PHA would not identify main-
Med high
RCM tenance strategies or detailed corro-
Med sion mitigation or identification strate-
Med low gies, it focuses on safety and not unit
reliability. In the end, the major deliv-
Low
erable is a set of actions that have to
Negligible be closed out to ensure compliance
Likelihood of failure (failure rate) with the PSM standard. Typically, this
Figure 1. This graphical “consequence-of-failure” risk matrix shows the areas covered by process haz- process is owned and facilitated by
ard analysis (PHA), risk-based inspection (RBI) and reliability centered maintenance (RCM) the PSM manager or department.
RBI. RBI arose from an industry
risk in a facility so that appropriate Let’s briefly characterize each. study in the 1990s that produced
and targeted mitigation actions can PHA. The PHA came out of OSHA’s API (American Petroleum Institute)
be taken at the right time. That being PSM standard and is one of the 14 580 and 581, which describe a sys-
said, there are three common risk elements listed above. Every five tematic risk identification and mitiga-
identification and mitigation frame- years, subject matter experts come tion framework that focuses only on
works that are currently being used together for a couple of weeks and loss of containment. For this reason,
in process facilities today. These in- identify the major events that could when an equipment item or pip-
clude process hazard analysis (PHA), happen at different “nodes” in a unit. ing segment (typically called “piping
risk-based inspection (RBI), and reli- The general idea is to use guide- circuit”) is evaluated, the only failure
ability-centered maintenance (RCM). words to systematically focus the that is of concern to the facility is the
QuestIntegrity.com
CHALLENGE CONVENTION
Circle 30 on p. 94 or go to adlinks.chemengonline.com/61495-30
CE-half-page-FTIS-Feb-2016.indd 1 1/15/2016 2:10:50 PM
HEXOLOY
®
The Sustainable
Solution in the Harshest Environments
■ Universal corrosion resistance in extremely high
temperatures; small to large complex shape & tight
tolerance capabilities.
■ Our customers minimize maintenance, energy
& downtime costs, reduce emissions & maximize
productivity through Saint-Gobain Engineered
Ceramics' innovation & relentless sustainability focus.
Leverage Saint-Gobain's
Corrosion & Heat Problems that
■
Circle 34 on p. 94 or go to adlinks.chemengonline.com/61495-34
I
n the chemical process industries
(CPI), one incident can have a tre-
mendous impact on the people
in the plant, the communities
around it, the environment and the
production asset.
This article outlines how learn-
ing from past incidents continues
to drive the development of both
newer standards, as well as new ap-
proaches to process automation as
it relates to plant safety and security.
many cases, safety is not the elimi- Connecting PSM and FS safety instrumented systems (SIS)
nation of risk, which could be im- Organizations, such as OSHA, rec- that were designed following a pre-
practical or unfeasible. ognize Functional Safety Standard vious RAGAGEP, and to effectively
Although the CPI must accept ISA 84 as a Recognized and Gen- keep its older equipment as long
some degree of risk, that risk needs erally Accepted Good Engineering as the company has determined
to be managed to an acceptable Practice (RAGAGEP) and one way that the equipment is designed,
level; which in turn makes safety a to meet the PSM requirements de- maintained, inspected, tested and
societal term as well as an engineer- fined in 29 CFR 1910.199. Apply- operated in a safe manner. As in-
ing term. Society establishes what ing ISA 84 is more than purchasing dicated by Klein [3], that does not
is commonly accepted as safe and a technology with a given certifica- mean that the existing system can
engineers have to manage risk by tion or using a particular technology be grandfathered and ignored from
introducing risk-reduction methods scheme or architecture. Industry best that point forward.
including human elements, such practices such as ISA 84 consider a The intent of the clause is for
as company culture and work pro- great deal of applied learning. ISA the user to determine if the PSM-
cesses and technologies that make 84 is a performance-based standard covered equipment, which was de-
the production facilities an accept- and describes multiple steps before signed and constructed to comply
able place to work and a responsible and after selecting and implementing with codes, standards or practices
neighbor in our communities. safety system technologies. These no longer in general use, can con-
The CPI has applied learnings steps — commonly referred to as tinue to operate in a safe manner,
from numerous events over the last the safety lifecycle — are also the re- and to document the findings.
40 years. These incidents and ac- sult of applying lessons learned from Therefore, the emphasis should be
cidents have resulted in changes to incidents and events. on the second part of the clause,
regulations and legislation and have Research (as documented in the which states that “the owner/op-
driven the adoption of best practices book “Out of Control” [2]) has shown erator shall determine that the
that address the known factors at that many industrial accidents have equipment is designed, main-
the root of those events. their root cause in poor specifica- tained, inspected, tested and op-
A lot of the best practices are re- tion or inadequate design (about erated in a safe manner.” And that
lated to understanding and evalu- 58%). Additionally, users should determination is a continuous effort
ating hazards and defining the ap- consider that installing a system is that should be periodically revised
propriate risk reduction, including not the “end of the road,” but rather until said equipment is removed
measuring the effectiveness of the another step in the lifecycle of the from operation and replaced with a
methodologies or technologies used facility. Approximately 21% of inci- system that is designed in line with
in reducing the risk. dents are associated with changes current best practices.
Risk-reduction methods using after the process is running, and Another consideration is that the
technology — including digital sys- about 15% occur during operation clause would cover not only hardware
tems — have received extensive cov- and maintenance. and software, but also management
erage in trade publications over time and documentation, including mainte-
as they are important contributors to ISA 84’s grandfather clause nance, all of which should follow cur-
process safety and plant productiv- It is well-known that Functional rent standards — that is, the most re-
ity. However, it is critical to recognize Safety Standard ISA 84.01-2004 cent version of ISA 84 or IEC 61511.
human factors and their impact on contains a grandfather clause
process safety in the design, selec- based on OSHA regulation Emerging technologies
tion, implementation and operation 1910.119. This clause allows users The last few decades have seen
of technology. to continue the use of pre-existing technology changing in all aspects
D
quantity or quality of product(s).
uring the design of a chemi- From a process standpoint, a Utility
cal process plant, the main chemical process plant is a combi-
focus is on which process nation of equipment, utility networks
units or unit operations and control systems. To design a Control valve, instruments
must be integrated to convert the plant with sufficient flexibility, each FIGURE 1. Different elements of a plant need dif-
feed streams into product stream(s). of these three elements needs to ferent levels of operating flexibility. Since the utility
Design engineers work to achieve allow flexibility. Generally speaking, network provides support duty to the equipment,
this goal; however, in terms of mak- the control system (including con- it needs a higher turndown ratio. Control valves
ing sure the plant operates smoothly, trol valves and sensors) and util- and other instruments have a duty to take care of
equipment across a wider operating range; thus
which is equally important for opera- ity network should offer the largest they require an even higher rangeability
tion engineers and operators, there amount of operating flexibility, while
are less well-know parameters fac- the equipment itself could offer the or even to support the shutdown of
ing the design engineers. lowest amount of flexibility (Figure 1). downstream equipment. But it may
There are five primary process This requirement for larger flexibility also occur accidentally due to, for
parameters in each plant — flow, for control items and utility network example, a drop in feed flowrate.
(liquid) level, pressure, temperature, considerations is important because But process plant operators like
and composition. Composition can of the supporting role of the util- to know by how much the flowrate
be considered a collective term that ity system and the controlling role of the equipment (and in the larger
reflects all parameters (chemical and played by instruments in a plant. sense, the entire plant capacity) can
physical), and provides an indicator Two important concepts are used be decreased without compromis-
of the quality of the stream. Com- to quantify flexibility: turndown (TD) ing the process goal or generating
position can be used to describe ratio and rangeability. These are off-specification product. Thus, TD
the moisture of a gas stream or the discussed below, and illustrated in ratio can be defined as the ratio of
octane number of a gasoline stream, Figure 2. high flow to normal flow, as shown in
or even the electric conductivity of a Equation 1.
water stream. Turndown ratio
During operation, equipment pro- The flexibility of equipment or a plant QHigh
cess parameters generally deviate can be defined using the TD ratio. TD ratio =
from the design values (normal level) The most common definition for TD
QLow (1)
over time. Five levels can be defined ratio is “ratio of the normal maximum
for each process parameter: nor- parameter (numerator) to the normal QHigh = the flowrate of the system at
mal level, high level, high-high level, minimum parameter (denominator).” high level
low level and low-low level. In es- However, the meaning of “normal QLow = the flowrate at low level
sence, the operational parameters maximum parameter” and “normal The numerical value of the TD ratio
of a plant relate to the behavior of minimum parameter” is not always is typically reported as a ratio, such
the plant between the low level and clear and the interpretation may vary as 2:1.
high level of each parameter of the in different companies and plants It is important to note that the
individual equipment components, (This is discussed below). denominator term is flowrate in low
individual units or the entire plant. For an individual equipment com- level, and not low-low level. This is
In most cases, the operability of a ponent, or multi-component equip- important as it is the differentiator
plant can be defined using at least ment systems, low-flow or low-ca- between the concept of TD ratio
three key parameters: flexibility in pacity operation happens frequently and rangeability, which is discussed
operation, resistance against surge over the lifetime of a plant. The re- later. Generally, flowrate in low level
(or upset) and the speed of recovery duced-capacity operation may be (as shown in Figure 2) is considered
from upset. intentional or accidental. to be the minimum level of flow at
For instance, reduced-capacity which the process goals can still
Maintaining operating flexibility operation could be planned for the be reached.
Flexibility of operation in this context purpose of off-loading the equipment However, there is another interpre-
means the ability of a plant to operate for inspection, testing, monitoring, tation of TD ratio that is often used
72 CHEMICAL ENGINEERING WWW.CHEMENGONLINE.COM SEPTEMBER 2015
TABLE 1. TURNDOWN RATIO OF SELECT EQUIPMENT 2:1 for a plant. The required TD
Item Turndown ratio ratio could be as high as 3:1 or 4:1
Pipe Large, but depends on the definition of maximum and minimum flow
for a plant.
Storage containers Very large; The maximum value is the total volume of the container, but the
(Tank or vessels) minimum value could be dictated by a downstream component. For example,
Equipment flexibility
a centrifugal pump may dictate a minimum volume to provide required NPSH The TD ratio can also be determined
Centrifugal pump Typically: 3:1 to 5:1
for a given piece of equipment, using
other values that are stated for the
Positive-displ. pump Theoretically infinite
component. For example, even when
Heat exchanger Small, depends on the type; for instance, less than 1.5:1 a TD ratio is not explicitly stated for a
Burner [1] Depends on the type; for example: centrifugal pump, when the pump is
Pressure jet type: ≈ 2:1 said to have a capacity of 100 m3/h
Twin fluid-atomizing type: >8:1
and a minimum flow of 30 m3/h, this
TABLE 2. UTILITY SURGE CONTAINER TO PROVIDE TD RATIO means that the centrifugal pump has
a TD ratio of 3:1.
Surge container Residence time
The TD ratio of a reciprocating
Instrument air (IA) Air receiver 5–10 min. or higher depending on pump could theoretically be defined
whether it is connected to UA or not
as infinite because it can work over a
Utility water (UW) Water tank Several hours very wide range of flows. However, in
Utility steam (US) Utility steam cannot be stored for a long time without condensing; the options practice, such a pump cannot handle
for storing steam are the steam drum of a boiler, or if a conventional boiler is any flowrate that fails to fill the cylin-
not available, a vessel as an “external steam drum” could do the same task
der of the pump in one stroke. Partial
Utility air (UA) No dedicated container; could “float” with IA filling of the cylinder may cause some
Cooling water (CW) Cooling tower basin Depends on the size of the network damage to mechanical components
Cooling/heating glycol Expansion drum Depends on the size of the network of the pump over the long term. Thus
the minimum required flow is a func-
TABLE 3. TURNDOWN RATIO OF SELECT INSTRUMENTS tion of cylinder volume and stroke
Item Turndown ratio speed of a specific pump.
Flowmeter: orifice-type 3:1 [2 ] The TD ratio for pipelines presents
Flowmeter: vortex-type 10:1 to 50:1 [2 ] a more complicated situation. With
piping systems, there are several dif-
Flowmeter: Coriolis-type 5:1 to 25:1 [2 ]
ferent ways to define the minimum
Control valve Depends on type and characteristics; generally 50:1, and less than 100:1 flow. For instance, it could be defined
TABLE 4. ARBITRARY VALUES OF FLEXIBILITY PARAMETERS as the minimum flow that does not
Low flexibility Medium flexibility High flexibility
fall into the laminar flow regime. Or,
it could be considered as the mini-
Equipment (TD ratio) < 1.2:1 to 2:1 2:1 to 3:1 5:1 to 8:1
mum flow that keeps a check valve
Instrument, control valves (rangeability) ≈ 4:1 10:1 to 30:1 20:1 to 100:1 open (if a check valve is used).
For liquid flows in pipes, the mini-
by operations staff. During opera- number. In general, the academic mum flow is more commonly inter-
tion, people expect the TD ratio to definition of TD ratio generally uses preted as the minimum flow that
answer the question in this scenario: a high-to-low values set up, while in makes the pipe full, or the sealing
“My plant is running normally and the field, operators often define TD flowrate (that is, no partial flow), or a
all parameters are normal. How- ratio using normal-to-low values. flow threshold below which the fluid
ever, occasionally, because of differ- The TD ratio can be defined for will freeze in an outdoor pipe. If the
ent reasons (including shortage of parameters other than flowrate, but flow bears suspended solids, the
feed, reduced plant or unit capac- it generally refers to flowrate. One minimum flowrate could be defined
ity), the flowrate falls. What is the reason for this is because flowrate as that at which sedimentation of
minimum value I can withstand can be the most important param- suspended solids may occur.
without compromising the quality of eter of a plant, helping to define the Table 1 provides examples of
the product?” economy of the system. The other typical values and rules of thumb
They basically interpret the TD ratio reason is because the flowrate might regarding the TD ratio for various
so that the numerator is the “normal be influenced by constraints out- types of process equipment. Note
level parameter” (and not the “high side of the plant (for instance, a lack that in Table 1, the TD ratio of stor-
level parameter”). However, the dif- of stored feed), which the control age containers is relatively large. This
ference in the interpretation does not system cannot necessarily adjust high TD helps to explain why large
generate a big difference in numerical (thus making a reduction in flowrate containers are used for surge damp-
value of TD ratio, as the normal and unavoidable). ening as part of a typical plant-wide
high level of parameters are often While the TD ratio is not always a control system.
not very far from each other. Due to requested parameter, and is often In some cases deciding on a re-
this potential confusion, the TD ratio not mentioned in project documents quired TD ratio needs good judg-
should be considered an approxi- for design purposes, operators are ment. One example is chemical-
mate parameter and not a precise usually looking for a TD ratio of least injection packages. The TD ratio
Turndown ratio
Rangeability
Vs.
Recirculation pump
Compressor
Recirculation pipe
Unit
Feed
FIGURE 4. By providing a recirculation pipe, the
turndown ratio of a piece of equipment can be Distillate
increased. If the fluid pressure is not enough, a
pump (or compressor) may be needed, and a con-
trol system is definitely needed
0
BATCH 1 BATCH 2 BATCH 3
Circle 24 on p. 94 or go to adlinks.chemengonline.com/56201-24
m • Social Media
TM
Circle 27 on p. 94 or go to adlinks.chemengonline.com/56201-27
78 CHEMICAL ENGINEERING WWW.CHEMENGONLINE.COM SEPTEMBER 2015
Solids Processing
Flowing core
Static
material
Robert McGregor FIGURE 1. Three common types of flow behavior for powder in a bin are mass flow (1a), core flow or fun-
Brookfield Engineering Laboratories nel flow (1b) and rathole formation (1c)
P
have been used in the minerals in- from a container. Figure 1a shows
owder jams are the once- dustry for decades. Recent improve- how particles move uniformly down-
in-a-month catastrophe that ments in the design of this equipment ward in lockstep with one another as
can bring processing opera- and the processing power available the fill level in the bin reduces. The
tions to a standstill. Whether in today’s personal computers (PCs) fundamental principle is referred to
it’s erratic flow behavior or complete make them more affordable and user as “first in, first out.” One obvious
stoppage of powder discharge, the friendly. The bottom line is that shear advantage is that blends of powders
consequence is the same. Shutdown cells can predict powder flow behav- retain their component ratio without
may be necessary before startup can ior using a proven scientific principle segregation. This is one of the most
take place. Why? Formulations often that measures inter-particle sliding important considerations for formu-
involve multiple component powders friction. Mathematical calculations lators who must ensure that final
blended together. If the flow becomes embedded in the software used product has the intended makeup
disrupted, one of the possible conse- with shear cells provide estimates as designed in research and devel-
quences is segregation of compo- for “arching dimension” in mass flow opment (R&D).
nents. Smooth and continuous flow and “rathole diameter” in core flow. More typical of powder process-
of powder from start to finish is the These values become design limits ing in most plant operations is “core
operating goal to minimize the onset for hopper openings and half angle. flow” or “funnel flow” as shown in
of other problems like segregation. This article addresses the rheol- Figure 1b. Particles at the top of the
Traditional testing techniques ogy of powder-flow behavior and container move toward the center
used to predict flow performance, explains how the shear cell is used and then downward through the
such as flow cup, angle-of-repose to make these types of powder mea- middle, discharging out the hopper
measurement and tap test, actually surements and calculations for stor- well before the powder that had been
have limited relevance to whether a age equipment design (see also, “A further down in the vessel. Larger
powder will flow. They are relatively pragmatic Approach to Powder Pro- particles have a tendency to move
affordable in terms of equipment cessing,” Chem. Eng., August 2015, more readily than smaller particles,
purchase and easy for operators to pp. 59–62). potentially resulting in segregation.
use. The data, however, do not pre- This type of behavior is called “last in,
dict whether reliable discharge will Types of powder flow first out.” One possible unfortunate
take place from the storage vessels In a perfect world for powder proces- consequence is that powder around
containing the powder. sors, “mass flow” would take place the outer wall of the vessel becomes
Shear cells for testing powder flow all the time when powder discharges stagnant, consolidates over time,
62 CHEMICAL ENGINEERING WWW.CHEMENGONLINE.COM SEPTEMBER 2015
is needed to allow the powder to dis- sample. The “tap density,” ρtapped,
charge from the cup. In a practical is calculated by dividing the reduced
sense, this instrument is used as a volume of powder at the end of the
“go” or “no-go” indicator for powder test into the sample weight. The two
processing on a regular basis. density values are compared to one
Angle of repose. This is a simple another, giving an indicator for the
test method that observes powder in consolidation that can take place
a pile and measures the angle of the over time when the powder settles.
pile relative to horizontal. Note that Two standard calculations that are
both the angle-of-repose method typically used by industry to evaluate
and the flow-cup test work with pow- tap test data are called Carr Index
ders that are loosely consolidated. (Carr%) and Hausner Ratio (HR), as
They do not attempt to evaluate the defined in Equations (1) and (2):
powder as it settles, which is what
happens when powder is placed in a
containment vessel of any kind. This
phenomenon, called “consolidation,”
is an important distinction to keep in (1)
mind because it has direct impact on
FIGURE 2. The flow cup test is relatively easy to how flow behavior can change.
setup and perform, and the data are used to cal-
culate the Carr index, Equations (1), and Hausner
Tap test. The tap test takes a cylin- (2)
ration, Equation (2) der of powder and shakes it to de-
termine how much settling will occur.
and then becomes lodged in place. The change in volume of the powder Shear cell test for flowability
This type of structure is referred to as from start to finish is a measurement Shear cells measure the inter-parti-
a “rathole” shown in Figure 1c. The of the powder’s tendency to consoli- cle friction of powder materials. This
rathole may extend from top to bot- date. The “loose fill” density, ρpoured, type of test has direct application
tom of the bin and may change in of the powder at the start of the test to predicting flow behavior in grav-
diameter of opening as a function of is calculated by dividing the cylin- ity discharge for powders stored in
powder depth. der volume into the weight of the vessels of any kind. Shear cells were
Processors prefer mass flow for
obvious reasons. Cohesive materi-
als will generally exhibit core flow
in plant equipment as originally de-
signed. The hopper wall angle and
its material of construction have a di- ENSURE
rect impact on flow behavior. There-
fore the challenge is to manage the YOUR PIPING
problem with the equipment that
exists, which means modifying the INTEGRITY
formulation, or redesigning the bin
equipment, if practical.
Cohesive
onto a test instrument such as that 6.0
Easy flowing
shown in Figure 3b. The lid, which Free flowing
5.0 Non flowing Data set #1
will fit on top of the cell, is attached Data set #2
4.0 Cohesive Data set #3
to the upper plate on the instrument Data set #4
and can be one of two types: 3.0
1. The vane lid (Figure 3c) has individ- Easy flowing
2.0
ual pockets separated by vanes.
2. The wall-friction lid (Figure 3d) is a flat 1.0
surface and is made of material sim- 0.0
Free flowing
ilar to the hopper wall in the powder
0 1 2 3 4 5 6 7 8 9 10 11 12 13
storage vessel on the production Major principal consolidating stress, kPa
floor. Examples might include mild FIGURE 4. The flow-function graph shows how the failure strength for the powder changes as a function
steel, stainless steel or Tivar. of increasing consolidating stress
800
Concluding remarks
600
Shear cells provide a scientific basis
400
for analytically predicting flowability
of powder in gravity discharge. Their
200 use is becoming more accepted
because improved designs for the
0 instrument make them affordable,
0 1 2 3 4 5 6 7 8 9 10 11 12 13
Major principal consolidating stress, kPa user friendly, and automatic in op-
FIGURE 6. The density of a powder in a vessel will vary depending on the consolidating stress, which in eration under control of a computer.
turn is a function of the fill level The most notable change in the past
year is the reduction in time needed
ing data form the flow-function graph cern. Data from this test may also to run a standard flow-function test
(Figure 4), which shows how the fail- have some correlation with findings from 45 min to 15 min. Productivity
ure strength for the powder changes obtained in the angle-of-repose test gains with the current generation of
as a function of increasing consolidat- described earlier in this article. instrumentation certainly give rise to
ing stress (height of powder-fill level Density. Density of powder in a ves- their potential use in quality control
in the vessel). Industry has agreed sel will vary depending on the con- as well as R&D. The chemical pro-
to classify regions of flow behavior solidating stress, which in turn is cess industries on the whole view
as shown in the figure, ranging from a function of the fill level. Figure 6 the shear cell as an important tool
“free flowing” to “non-flowing.” As shows an example. If the change in for improving rapid scaleup of new
might be expected, many powders density increases by more than 50% formulations into full production. n
exhibit “cohesive” or “very cohesive” relative to the “loose fill” condition, Edited by Gerald Ondrey
flow and are likely to be problematical then there is an expectation that
in terms of processability. flow problems may exist. Note that Acknowledgement
Wall friction. The wall-friction test the density test will very likely have All photos courtesy of Brookfield En-
measures the flowability of the pow- a point on the curve that correlates gineering Laboratories, Inc.
der on the material comprising the with the findings in the tap test de-
hopper wall. Data from the wall- scribed earlier in this article. Author
friction test (Figure 5) show how the Robert McGregor is the general
effective friction angle for the hopper Data analysis manager, global marketing and
sales for High-End Laboratory In-
wall to allow gravity-driven powder Parameters of interest that can be struments at Brookfield Engineer-
flow on its surface changes as a func- calculated from the data in the above ing Laboratories, Inc. (11 Com-
tion of consolidating stress (height of tests include the following: merce Blvd., Middleboro, MA
powder-fill level in the vessel). Expe- 1. The arching dimension is the 02346; Phone: 508-946-6200
ext 7143; Email: r_mcgregor@
rience indicates that friction angles length of a bridge section that the brookfieldengineering.com; Web:
below 15 deg will have relatively easy powder has sufficient strength to www.brookfieldengineering.com).
flow behavior whereas friction angles create in the hopper section of a He holds M.S. and B.S. degrees in mechanical engi-
neering from MIT (Cambridge, Mass.; www.mit.edu).
above 30 deg will be cause for con- vessel. If the bridge is longer than
Dechlorination
IN BRIEF Pump Bar screen
Chlorine
contact
T
here is growing interest in auto- FIGURE 1. Most wastewater treatment systems use a com-
mating wastewater treatment pro- mon sequence of steps to treat influent wastewater and then
discharge, store or reuse it in line with local regulations.
cesses across a broad range of Automating this approach helps an operator more effectively
industries. In particular, a paradigm manage and treat wastewater, saving time and money in the
shift is starting in automating industrial process
wastewater treatment in various sectors of
the chemical process industries (CPI), such undergo tertiary treatments to be further oxi-
as foods (especially grain processing, sug- dized or disinfected, or to undergo additional
ars, sweeteners and edible oils), beverages purification, including by granular activated
(mainly soft drink bottlers and breweries), carbon (GAC) or membrane separation, be-
and hydrocarbon and chemical processing fore reuse or discharge to a public sewer or
(particularly petroleum and petrochemical open body of water.
plants). The driving forces behind this evo- A fully optimized, industrial wastewater-
lution are economic. Wastewater process treatment plant will operate at a lower total
optimization most often leads to a more cost of materials, labor and energy to do the
efficient use of chemicals, reduced energy following:
consumption and less solid waste. • Remove or reduce large solids and par-
Most wastewater-treatment systems use a ticulate matter (primary)
common sequence of steps (Figure 1), with • Remove or reduce fats, free oil (and
the purpose of first removing solids materi- grease), dispersed oil and emulsions
als in the influent wastewater, recovering • Remove organic materials efficiently
lost product, removing solids, fats, oils and (secondary) and withstand higher vari-
greases (FAG), treating the water biologi- able loading, with enhanced, biological
cally and chemically enhancing flocculation, activated sludge systems through:
coagulation and physical removal of the bio- ❍❍ Control of dissolved oxygen levels,
logical solids and sludge. The clarified and minimizing energy required for
decanted wastewater is the effluent that may aeration
44 CHEMICAL ENGINEERING WWW.CHEMENGONLINE.COM SEPTEMBER 2016
Circle 22 on p. 98 or go to adlinks.chemengonline.com/61499-22
❍❍Maintaining food-to-mass ra- • Disinfect pathogens and produce pending upon the reuse application
tio, pH and nutrient balance, effluent water quality for reuse or and corresponding water quality re-
minimizing chemical usage below discharge limits to open quirements, tertiary disinfection for
and system upsets body, waterway or public waste- pathogens and final polishing with
• Produce a readily settleable bio- water treatment plant GAC or reverse osmosis (or both)
logical floc (small microbial mass); More advanced integration of may be needed.
less energy to coagulate and technologies can be applied to meet
separate (Figure 2) requirements for reuse, whether Implementing process control
• Generate minimal volume of within the facility (for example, wash In general industry, process auto-
sludge and biosolids to dewater, water), for irrigation and agricultural mation is ubiquitous and integral to
minimizing energy, chemical us- purposes or higher purity applica- upstream control mechanisms and
age and disposal costs tions, like clean water utilities. De- production yield. Statistical process
control (SPC) can use process ana-
lytical technology to generate high-
value data in real- and near-time,
and is critical to closely control pro-
cesses, quality and maximum pro-
duction yield. There is a prevailing
interest across industries to identify
opportunities to gain process knowl-
edge by understanding process ef-
fluent streams. These waste streams
Your Valve and combine to become the wastewater
treatment influent. Companies are
Instrumentation investing in multiple tools, devices,
analyzers and sensors, and integrat-
Partner ing these measurements into pro-
cess automation and control sys-
tems for the wastewater treatment
GEMÜ Valves features plant (WWTP). They are looking at
collecting useful data with the right
diaphragm valves, parameters, and applying SPC tools,
previously reserved for production
butterfly valves, angle purposes, to continually analyze and
seat and globe valves, optimize their wastewater treatment
processes. The proper design and
lined metal valves, execution of experiments can help
show the pertinent relationships be-
flow measurement, tween multiple parameters that yield
and multiport the best process performance. The
application of this empirical process
block valves. knowledge can translate into sig-
nificant performance improvements
and efficiencies.
Process parameters
Depending upon the physical and
chemical characteristics of waste
Visit us at streams, a number of treatment
modules are employed to remove,
Booth #134 reduce and change sample stream
constituents including, but not lim-
ited to, the following:
3800 Camp Creek Parkway • Building 2600 • Suite 120
• Bar screens and strainers for grit
Atlanta, GA 30331 • 678-553-3400 • info@gemu.com
and particles
• API (American Petroleum Institute)
www.gemu.com separators and corrugated plate
separation for free oil and grease
• Chemicals and dissolved or in-
duced gas (or air) flotation for oily
Circle 17 on p. 98 or go to adlinks.chemengonline.com/61499-17
46
chemEng201609_gemu_isld.indd 1
CHEMICAL ENGINEERING
8/1/2016 3:12:07 PM
WWW.CHEMENGONLINE.COM SEPTEMBER 2016
solids and emulsified oils Organic carbon: The influent, or-
• Biological activated sludge and ganic carbon loading is a key pro-
advanced membranes for organ- cess parameter for a WWTP, and
ics, nitrogen and heavy metals has historically been quantified using
• Physical and chemical clarification chemical oxygen demand (2 hours)
and advanced membranes for or biochemical oxygen demand (5
microbial flocs days; BOD5). With the availability of
• Chlorine (gas, hypochlorite and online, process instrumentation for
chlorine dioxide solution) and total organic carbon (TOC) analysis,
ozone for trace organics and a direct measurement of the organic
pathogens
• Granular activated carbon (GAC)
concentration can be used to im-
prove downstream performance.
PROVEN
for organics
• Chemical disinfection for patho-
Specifically, by knowing the exact
values of TOC, the plant can be op-
PERFORMANCE
gens (typically chlorination) erated to accommodate variation in
• UV (ultraviolet) for pathogens, the amount of organics, and remove
trace organics and residual them efficiently. For instance, there
ozone destruction is often an introduction of chemicals
• Chemical pH neutralization (such as potassium permanganate,
• Reverse osmosis for inorganics hydrogen peroxide or chlorine) after ROTOFORM
and minerals primary solids removal to reduce the GRANULATION
By employing a combination of total oxygen demand, often referred
discrete (grab) and online mea- to as pre-oxidation. This step can FOR PETROCHEMICALS
surements before, after and at in- be eliminated with lower influent or- AND OLEOCHEMICALS
termediate process points, each ganic concentrations, or minimized
module’s performance can be by using it only when the load is
monitored and improved over time. above a threshold limit based on the
Some of the parameters measured plant’s treatment capacity.
by the available probes, meters, Dissolved oxygen: In a biological or
sensors and analyzers include: activated sludge system, there is an
flow, pH/ORP (oxidation-reduction opportunity to adjust the amount of
potential), conductivity, dissolved dissolved oxygen generated by the
oxygen (DO), suspended solids, aeration system to a level commen-
specific ions [for example: nitrogen surate with the organic load, while High productivity solidification of
(ammonia, nitrates, nitrites), phos- avoiding excessive aeration that can products as different as resins, hot
phorus (phosphates), chlorine], shear or tear the biological flocs, melts, waxes, fat chemicals and
total organic carbon, sludge den- which in turn reduces the overall ef- caprolactam has made Rotoform® the
sity index and turbidity. fectiveness of organics and biosol- granulation system of choice for
chemical processors the world over.
Free oil and grease: Before intro- ids removal. Continuous monitoring
Whatever your solidification
duction of the waste stream to the of influent organic loading and dis-
requirements, choose Rotoform for
biological or activated sludge sys- solved oxygen levels in select zones reliable, proven performance and a
tem, free oil and grease should be of the activated sludge basin provide premium quality end product.
removed or reduced to below a an opportunity to optimize the aera-
maximum threshold of 50 mg/L, and tion system, the largest energy ex- High productivity –
ideally below 25 mg/L, to avoid in- pense in the operation of a WWTP. on-stream factor of 96%
terfering with the microbial activity. Food-to-mass ratio: Industrial Proven Rotoform technology –
Some of the negative repercussions wastewater-treatment systems are nearly 2000 systems installed
of allowing excess levels of free oil looking at the ratio of organic load in 30+ years
Complete process lines or
to come into contact with the bio- or “food,” to the total biomass pres-
retrofit of existing equipment
mass are rapid oxygen depletion, ent in the biological system. The
Global service / spare parts supply
encapsulation of the bacteria, and biomass of the mixed liquor can be
foaming. Depending upon the levels estimated by measuring mixed li-
of free oil, and geometry of the oil quor suspended solids and sludge
droplets, one can use API separa- density. This F:M or food-to-mass
tors or corrugated-plate separation. ratio, is a critical process control
Dispersed and emulsified oils are re- parameter that can indicate system Sandvik Process Systems
moved and reduced through a com- overload or when there are insuffi- Division of Sandvik Materials Technology Deutschland GmbH
Salierstr. 35, 70736 Fellbach, Germany
bination of chemicals, for lowering cient organics to “feed” the micro- Tel: +49 711 5105-0 · Fax: +49 711 5105-152
info.spsde@sandvik.com
pH and enhancing the dissolved or bial population. The plant operation
www.processsystems.sandvik.com
induced gas flotation unit(s). can use near realtime information
Circle 38 on p. 98 or go to adlinks.chemengonline.com/61499-38
0
BATCH 1 BATCH 2 BATCH 3
Circle 23 on p. 98 or go to adlinks.chemengonline.com/61499-23
tion can be accomplished by several differ- tion can also be used to prevent or mitigate
ent chemical and physical methods, such as the negative impact of process upsets. In
chlorine gas, sodium or calcium hypochlo- the case of an unexpected event or exces-
rite solution, chlorine dioxide, ozone, and sive “shock” load to the system, the influent,
UV light (254 nanometer wavelength). After online TOC measurement can be used to
disinfection, the end-of-pipe purpose will de- automatically divert to an equalization basin
termine if additional treatment is necessary. or temporary storage vessel, sometimes re-
Some industrial utilities have reused waste- ferred to as a calamity tank.
water with a GAC step to absorb organics Effluent discharge monitoring: Meet-
and excess chlorine, and reverse-osmosis ing regulatory requirements for effluent dis-
membrane separation to remove inorganics charge levels is critical to any business op-
and trace organics, achieving higher purity. eration. There are continuous monitors for
Managing process upsets: Upsets in the many of the common effluent-wastewater-
wastewater process can affect removal effi- quality characteristics, including pH, dis-
ciencies at each treatment step. More severe solved oxygen, total dissolved solids, total
upsets can overload a system, even lead- suspended solids, and total organic carbon
ing to the loss of an entire, activated sludge (often used to trend chemical and biochemi-
biomass. The cost and time to reseed and cal oxygen demand). Finally, effluent pH for
restore lost biomass are significant, often discharge should almost always be neutral,
upwards of tens of thousands of dollars and ideally pH 6.8 –7.2.
several months. Real- and near-time detec- Solids disposal: The biosolids produced
Circle 36 on p. 98 or go to adlinks.chemengonline.com/61499-36
50 CHEMICAL ENGINEERING WWW.CHEMENGONLINE.COM SEPTEMBER 2016
from management of the activated can be impacted by changes in energy consumption, and solid waste
sludge volume in the aeration basins the upstream processes, as well as disposal, by maintaining the dynamic
and during clarification, are typi- seasonal variations in environmen- treatment system in an optimal opera-
cally dewatered using a belt press tal conditions such as ambient tem- tional state. n
or centrifuge, before being used as perature and rainfall. Edited by Dorothy Lozowski
fertilizer or disposed of as waste. By employing continuous process
The cost of sludge handling and de- monitoring tools and integration to Author
watering, in energy, chemical usage automation and process control sys- J.P. Pasterczyk is the corporate
and disposal, is often the second tems, more industries are finding bet- key accounts manager — analyti-
cal instruments for GE Water &
highest expense in a wastewater ter ways to effectively manage and Process Technologies (6060 Spine
treatment facility, after aeration. The treat their process and wastewater Road, Boulder, CO 80301-3687;
ability to use the dewatered sludge effluents. This automation provides Email: john.pasterczyk@ge.com;
Phone: 720-622-0166). He has
as fertilizer is dependent upon the more predictable and controllable pro- 25 years of international experi-
content of undesirable constituents, cesses, reducing the frequency of up- ence in water and wastewater
treatment, from water quality
such as heavy metals or residual sets and assuring a more consistent monitoring to pretreatment, biological treatment pro-
pathogens, including fecal coliforms effluent that meets discharge require- cesses and disinfection. Pasterczyk has spent the last
such as E. coli (Escherichia coli). ments. The efficiency of the biological 17 years with GE’s Analytical Instruments, primarily fo-
cused on total organic carbon analysis and integration
Instead of land application for ag- system to remove organics depends of water quality monitoring with process automation in
ricultural purposes, the solid waste upon the quality of the upstream pro- petroleum refining and petrochemicals, chemical, mu-
can be compacted or incinerated cesses — oil and grease and solids nicipal water, pharmaceutical and semiconductor indus-
tries. He is an expert in industrial wastewater treatment,
(or both) to reduce volume for dis- removal, and the controllable, ambi- applied statistics, statistical process control and optimi-
posal. A more sustainable approach ent conditions, such as dissolved zation, Lean Six Sigma methods and advanced quality
is sending the sludge to anaerobic oxygen, food-to-mass ratio and nutri- management systems. Pasterczyk received a B.S. de-
gree in physics from Drexel University and a Master of
digesters to produce methane gas, ent balance (CNP ratio). Utilization of Engineering degree from the Lockheed Martin Engi-
which can be fed to gas-fired tur- process analytical instrumentation and neering Management Program at the University of Colo-
bines to generate electricity. automation controls enables these rado, specializing in business performance excellence
and applied statistics/Six Sigma.
facilities to reduce total chemical and
Implementing process analytics
The data for each measured pa-
rameter can be tracked through
a data collection and visualization
system. A wide range of commer-
cially available software, as well as
discrete supervisory control and
data acquisition (SCADA) systems,
are employed by treatment facilities
to monitor critical and complemen-
tary water-quality characteristics.
With these tools, each treatment
module indicates the measured pa-
rameters before, during and after
treatment, while steady-state con-
ditions can be established to bet-
ter detect and anticipate upset and
sub-optimal conditions. Many pa-
rameters integrate into a feedback
or feed-forward loop for chemical
feed, becoming statistical process
control applications. New, multi-
variate relationships can be tested
and inferred through sound experi-
mental design and intrinsically valid,
statistical analyses. Good process
data leads to process understand-
ing and SPC brings and maintains
processes in control. Empirical evi-
dence can support or modify pre-
liminary assumptions and control
schemes. This acquired learning
Circle 06 on p. 98 or go to adlinks.chemengonline.com/61499-06