Professional Documents
Culture Documents
PVS2017
July 17-18, 2017, Silver Spring, Maryland
PVS2017-3512
ABSTRACT
What appears to be a simple question is often quite difficult to answer depending on the
quantity of flow; and size, type, and location of piping. Even the reason for asking the question
can be varied and complex - ranging from environmental regulation, investment decisions,
aging infrastructure improvement planning, and new equipment evaluation. Absolute field
performance testing of power plant equipment yields valuable data that can be used in a variety
of ways.
National and International codes list several methods to measure water flow in a performance
application and provide realistic uncertainty estimates. Codes and standards exist for
equipment evaluation and contractual performance tests. These Codes, though, are
sometimes viewed as costly or perceived to impose additional risk on suppliers. Herein, we will
present how to obtain performance test data and how that data can be used.
In situ pump performance testing is often a difficult and sometimes costly endeavor mostly due
to the measurement of flow. The quantity of flow; and size, type, and location of piping all
dictate the feasibility of measuring flow and the level of uncertainty that may be obtained.
Numerous techniques and technologies exist that have been used successfully for decades,
each having positive and negative aspects to their use and viability. Their cost varies
significantly depending on the technology and intended use (i.e. permanent versus temporary).
The following information is intended to provide the reader with a sense of how pump
performance can change over time, pump testing/flow measurement techniques, and budgetary
prices associated with testing.
The simple answer to this question is “nothing”. Facilities have been using pump curves and
original design flow for decades. The design flow is often used for: reporting water usage to
environmental regulators, potential system modifications, and evaluating system performance.
However, over time all things change. Pump components wear and pipe can change due to
erosion and corrosion. In addition, environmental changes can occur that can change water
levels, water chemistry, and flow approach conditions. Facilities sometimes replace flow related
equipment or change operations which can change the flow from the original design.
These changes can modify either the original pump curve or system curve. Pump wear can
cause a pump curve to be lower than expected, and in the case of a pump rebuild with different
components, higher than expected. Similarly, environmental and system changes may cause
the system curve to be above or below the original design. See Figure 1 for an example of an
original pump and system curve and Figure 2 for potential pump and system curves.
The data in Table 1 presents the results of pump flow data recorded in the field versus original
design (see Table 1). As shown, results can vary from slightly higher than design to as much as
27% below design. Note that pump 16 is 2% over design because of an incorrect impeller
replacement during a rebuild. For reasons of confidentiality, the facility details cannot be
presented.
Dye-Dilution – The dye dilution method for measuring flow is a method that allows an
instantaneous flow to be measured by determining the dilution of a tracer injected into a flow.
The dye dilution method is based on a mass balance calculation. A small quantity of
fluorescent dye (typically Rhodamine WT) at high concentration is continuously injected at a
measured, constant rate into the test flow. Concentration of the fully mixed flow is determined
by fluorescence intensity measurements. The ratio of the injected concentration to the final
concentration, minus any background concentration in the incoming flow, multiplied by the
Area Velocity by Pitot – Flow can be determined by integrating point velocities measured by
Pitot probe, which measure velocity by simultaneously sensing impact and static pressure. This
technique is typically only feasible if special pipe penetrations (see Figure 5) are installed in
existing piping. The expense of adding wet tap type fittings to existing piping is often cost
prohibitive and sometimes physically unfeasible. In addition to pipe access, this technique also
requires a custom built and calibrated pitot probe, and an accurate measurement of the inside
diameter of pipe. Once diameter measurements are obtained, careful calculation of equal
areas and probe immersion depth are required. The velocities can only be measured along the
axes of the ports.
Area-Velocity by Current Meters – For large pump intakes, velocity integration using current
meters (see Figure 6) may be a viable flow measurement technology. Flow is measured by
integrating velocities and the cross sectional area. Typically a custom built site-specific frame
and meter racks are constructed for each application (see Figure 7). The meter rack is lowered
into the intake gate slot. Current meter velocity integration is a relatively slow measurement
with a single flow condition requiring 1.5-2 hours; however, there is no impact to plant operation
(i.e. no dewatering to implement).
Ultrasonic – Ultrasonic flow meters have been used in both temporary (clamp on) and
permanent (internal) installations. Ultrasonic pulse transit times are altered by the velocity of
the flowing fluid. The effect on this transit time provides the velocity in the pipe at the axis of
the meter transducers. Multiple transducers paths are employed to register the entire cross
sectional velocity. Accurate measurements of the transducer locations and of the pipe or
conduit dimensions are required. The measured velocities are integrated over the pipe area to
calculate the fluid flow. Ultrasonic technology is suitable for long and short pipe runs and
provides a relatively fast measurement. Permanent installations require site-specific
installations and do impact plant operations (i.e. dewatering) to install. Once installed, future
testing is fast and easy.
Additional Measurements – Collecting the following information is recommended. This data can
be invaluable for use after the initial flow measurement; when correlated to the measured flow
they can be used for later indication of the flow performance.
Pump inlet pressure
Gauged or water level
Pump outlet pressure
Pump speed
Additional differential pressure measurements in the system
Power
METHODS APPLIED
Referring to Table 1, the majority of the testing was performed using the dye dilution method
(see Table 2). The method is favored due to low cost, minimal disruption to facility operation,
and brief testing time as compared to the other methods described. The two techniques
previously described (Area Velocity and Ultrasonic), though not represented in Table 2, have
been presented herein due their use predominately in the Hydropower industry, and their
potential applicability in thermal power production.
Referring again to Table 1, the associated price of testing for the pumps listed is presented in
Table 3. The reader will notice that the pump tests on the upper half of the table are
significantly higher priced than the lower half of the table. This disparity is due to the testing at
nuclear facilities versus testing at fossil/biofuel facilities. Nuclear facilities require significantly
more effort in the aspects of training, mobilization, demobilization, and paperwork.
All of the flow measurement techniques/technologies listed in this paper have uncertainties on
the order of 2%. Flow measurement uncertainties consist of bias and precision related to:
Mixing (dye dilution)
Dead zones (dye dilution)
Injection flow (dye dilution)
Calibration (all)
Data acquisition and reduction (all)
Length measurement (area velocity)
See the REFERENCE section at the end of this paper for the codes that are primarily used for
pump testing and flow measurement.
CONCLUSION
Measuring pump flow rate is often a difficult task but not impossible. Multiple code accepted
techniques and technologies exist that have been in use for decades. Understanding flow
conveyance system details and the limitations of measurement options are keys to determining
the best approach. Historical flow data indicates that pump flow output may be off as much as
27% due to either pump degradation, environmental or system changes, or a combination of
these factors. The price of testing is not trivial, especially when dealing with the nuclear
industry requirements.
REFERENCES
American Society Mechanical Engineers (ASME). 2002. ASME PTC 18-2002, Hydraulic
Turbines and Pump-Turbines Performance Test Codes. Copyright 2002.
American Society Mechanical Engineers (ASME). 2005. ASME PTC 19.1-2005, Test
Uncertainty. Copyright 2005.
American Society Mechanical Engineers (ASME). 2004. ASME PTC 19.5-2004, Flow
Measurement. Copyright 2004.