Chapter 4 BALANCING DUCT SYSTEMS WITH DAMPERS 41 INTRODUCTION 4.8.1 Procedure for Basic Method As 42 BALANCING BY DESIGN 482 Example | Basic Method) 49 43. PROBLEMS WITH DAMPERS AND POTENTIAL 483 Advantages and Disadvantages of Basic SOLUTIONS a2 Method au 44 LIMITATIONS OF DAMPER EFFECTIVENESS ....43 49 INTERACTIONS BETWEEN THE SYSTEM AND 441 Characteristics of Blasigates 43 THE FAN ...... 3s an 45. STRATEGIC GOALS IN ADJUSTING DAMPERS ..45 49.1 Throttling the Fan Airflow 41 45. Overall Goal 492 Optismum Initial Fan Output An 452. Distribution Goal 49.3 Estimating Qa-deat 412 453 Total Airflow Goal 4.94 Fan Adjustment After Balancing, 413 454. System Pressure Goal 4.10 MODIFIED BASIC METHOD 413 46 DETERMINING INITIAL CONDITIONS 4.10.1 Procedure for Modified Basic Method .....4-13 461 Determining Airflow Goals for Each Hood 4-6 4.102. Example 2 (Modified Basic Method) 413 4.62. Determining Adjustment Goals During 4.10.3 Using Adjustment Targets That Are Not Equal Balancing 47 10 the Goal Values 415 463 Centerline Velocity Pressure Goals 47 4104 Branch Airflow Changes 416 4.64 Hood Static Pressure Goals 47 All TARGET METHOD +16 47 WHEN THE DISTRIBUTION SHOULD BE 4.1141 Procedure for Target Method “17 RE-BALANCED ar 4.112 Example 3 (Target Method) An? 47.1 Measures of Effectiveness of Damper 4.11.3 Advantages of Target Method 419 Adjustment 48 REFERENCES 419 48 BASIC AIR BALANCE METHOD : a8 Figure 4-1. Dlastgate Damper 44 Figure 44. Different Leading Edges as Figure 4.2 Flow Around a Damper 4-4 Figure 4-5 Example Duet System 46 Figure 4-3 Graph of Idol eik (1972) Values 4-4 Figure 4-6 Effect of Dampers on Fan Airflows 4-1242 Industrial Ventilation 44 INTRODUCTION Dampers are ventilation devices used to adjust the airflows through the branches in a duct system. A damper reduces the airflow to a given branch by adding to its resistance to flow. As its resistance to flow increases, airflow is diverted to altemate pathways in proportions that can be predicted mathematical- yn By judicious adjustment of all dampers in a system, a tech- nician can force the relative airflows through the branches 10 achieve a desired distribution. Ths is called “balancing” with dampers. If each branch receives the correct fraction ofthe fam airflow, then by changing the fan rotation rate one ean raise oF lower all branch airflows to the desired levels. Alternatively, ‘one can insert every damper atill further until the additional ‘cumulative resistance “chokes the fan down” to the desired level This chapter discusses how dampers work, when they should be used, when they should be adjusted, and how t0 adjust them. The chapter describes three methods of air bal- ancing with dampers. An example is included with each method as well as ¢ discussion of advantages and disadvan- tages. Each successive method uses less field measurement and adjustment but also requires more calculations. For simple systems of only a few branches, the Basic Method could be used. As the system becomes more complicated two more ‘methods are shown that require more computations but fewer re-adjustments of dampers. This could represent a time sw ings In balancing methods presented in this chapter, VPui(meas- ‘urement of Velocity Pressure at the center of duct) is some- times used for intermediate steps of measurement. This is done because it represents an approximate relative value when doing quick adjustments inthe setting of dampers. To be valid, a complete traverse must be performed to determine actual ar- flow but approximations may be usod for intermediate steps All initial and final readings must be done with complete tra- verse readings as recommended in thio Manual and the Ventilation Manual. The most accurate readings during the balance will also be accomplished by full traverses inthe into mediate steps of balancing but for lage systems, Va (if prop- sly stride be woe 4.2 BALANCING BY DESIGN As is shown in the Design chapter of the Ventilation Manual, it is quite possible o design systems that have distri butions that are acceptably close to the desired levels without employing dampers. This is done by judicious selection of components (e.2., elbows with higher or lower loss coet cients) and duct “sizes” (i.e, cross-sectional areas) as apart of normal duct system design. The advantage of systems designed to be balanced without dampers is that they should, keep the same distribution unless duct diameters or fittings are ‘moditied in some way. One can be reasonably certain that the distribution romaine at the lovel for which it was deoigned. With dampers, there is always the concern that someone will change the damper settings when they ahould not The primary disadvantage of systems balanced without dampers is the same as their primary advantage: the distribu- tion cannot be changed easily. If changing conditions or requirements necessitate a change in distribution, itis costly to achieve by substitutions of fittings with higher of lower loss coefficients or by changing duct diameters. Aside fom cost considerations, replacing fitings is not likely to add enough resistance to shift airflow substantially. Likewise, rebalancing systems by changing duet sizes generally requires replacement ‘of whole branches and submains. Such replacements are cost- ly and disruptive enough that managers are understandably reluctant to authorize the necessary expenditures, especially ‘when balancing with dampers is much ess expensive and dis- ruptive. In some cases (as specified by the National Fire Protection Association — NFPA), there may not be an option to use blast gates or dampers and the system must be installed using the Balance by Design Method. The designer must check NFPA requitements and other specifications to deter- ‘mine which method should be used 43 PROBLEMS WITH DAMPERS AND POTENTIAL, SOLUTIONS ‘Not all airflow distribution problems can be solved solely bby adjusting dampers, and there are many problems practition- ers might assoctate with dampers 1. The pressure required at the fan is substantially higher ifdampers are used instead of balance by design. This is true only ifthe fan speed is unnecessarily high and ‘one adjusts the dampers to “choke down” the fan air- flow without changing the fan speed. If dampers are ‘employed only to produce the desired distribution and the fan is adjusted property, the pressure requirement for the system will nearly always be lower than the same system balanced “by design.”*! However, the pressure requirements with or without dampers may be execssive ifthe airflow in any branch must be increased more than about 20-30% Ifa lange Jwrvase is ues, une slivukd Lousider increasing dhe duct diameter of the branch duct and at least the first subunain downstieaus of it Operators and others tend to open dampers on one branch without considering the effect on airflows in other branches. Since opening a damper on one branch increases its flow but reduces the airflow to all other branches, opening dampers can be a way of “stealing” additional airflow from everyone else. This can pro- duce a “race to the bottom” where all operators fully ‘open their dampers, making the dampers useless forall. “Thi problem can be avoided by locking the dampers (using padlocks, welding blades in place, ete). Ifthe‘contaminants appear to be poorly controlled, no securi- ty system will protect the dampers. Hence, itis critical to provide adequate control 3. Plugging is likely due to sticky or stringy contaminants ‘becoming caught om the dampers. This may be an insu- perable problem for some systems unless onc is willing to clean the dampers frequently, perhaps even daily, ‘This maintenance can be encouraged by connecting the damper fiting with easily removed clamps and by Tocating cleanouts near dampers. 4, Flammable materials caught on a damper could ereate ‘or exacerbate @ fire or explosion hazard. This may be ‘an insuperable problem for highly flammable contami 5. airflow is reduced using a damper, the velocity in some duets may fail befow the mintmum velocutes needed to prevent rapid settling, This can certainly happen If Uie duct Uiamneters are larger than needed, avoid this problem, one should either replace over: sized duc or increase the tant atllows 19 ensure su ficient duct velocities. Some authorities strongly urge “balance by design” instead of balancing with dampers because of the many disadvantages listed above (See also Ventilation Manual, Chapter 5), That may be a wise policy when first installing a new system, but the high direct and indirect costs of replacing system components to obtain anew balance “by design” may discourage the re-bal- ancing of systems, leading to prolonged periods of inadequate airflows to hoods. 4.4 LIMITATIONS OF DAMPER EFFECTIVENESS Care must be taken when attempting to make large changes to airflow distributions. There is no specific “limit,” but as dampers are inserted farther, the fan pressure requirement can increase. Furthermore, if the system carries particulates, care ‘must be taken to ensure that duct velocities do not fall below ‘minimum levels required to prevent plugging. ‘At the same time the placement of the damper is important, 1 proper operation. The blade should never be located at the bottom of the duct or near elbows and otner disturbances that ‘may influence the pressure drop and reliability. Damper blades can become sources for material buildup if not property iocat- ed. If'large changes in airflow to certain branches are required, it may be necessary to replace with either a larger or smaller diameter duct, 1. IF paniculates are involved and the airflow through a branch duct is to be reduced, it may be necessary 10 replace the duct with a smaller diameter in order to ‘maintain sufficiently high velocities atthe reduced air~ flow. Balancing Duct Systems with Dampers 4-3 2. If particulates are not involved, then airflows can be reduced without changing duct sizes. Unless a majori- ty of branches are virtually closed off 1s unikely that the fam pressure will rise to extreme levels (i.e, maxi- ‘mum fan pressure). Note that ran motor power require ‘ments will decline if centrifugal fans are used. 3. Increasing airflows to a given branch with dampers is ‘much more difficult unless one increases fan speed ‘The possible inerease to a given branch is limited by two factors: 1) Fan airflow falls as dampers are inser ced so that the amount available to other branches is somewhat reduced, and 2) The airflow shifted from an adjusted branch to other branches is shared by al other branches (but especially those just upstream). Hence, it is difficult to sharply increase the airflow to any given branch unless the airflows to almost all other branches, are reduced, As a rule of thumb, if the airflow through, a branch must be increased by more than about 20%, it is likely that damper adjustments will not be sufficient. IF large increases in airflow are retired for 9 given branch (especially ifthe other branch airlows are at desired levels), ane should consider replacing the branch duct witha larger size before adjusting dampers and the fan speed! AS n first appreximation, the new duct size can be determined using traditional methods ff duct sizing. That is, the dict area shone he the largest duct size whose area is less than the target air valine divided hy the ein ri air velocity 44.1 Characteristics of Blastgates. There arc many desigus aud varias Uf design UF adjustable dampers used i ‘exhaust ventilation systems, but the most common (and most sulwble for conauiiaan conuul ventilation) are “Vlsstgate cutoff” dampers (see Figure 4-1). The slide is inserted perpen- dicular to the flow. As the ais flows around the daunpet blade (see Figure 4-2), it separates from the duct downstream of the blade. Energy losses in the separation region vary with the fraction of the duet cross-section blocked by the damper, the sharpness of the damper, and the velocity pressure (VP) of the airflow upstream of the damper. It may also vary with the proximity of upstream and downstream disturbances, such as elbows, hood connections, etc. ‘These energy changes produce static pressure changes downstream of the damper. The resistance of the damper, ‘Xaanpes‘*"? can be determined from the change in total pressure due to the damper divided by the upstream velocity pressure Since itis difficult to measure pressure across a damper accu- rately, the most feasible way to determine the resistance of a damper is to determine the difference in the resistance at the end ofthe branch duct with and without the damper. Since the value of both the total pressure and velocity pressure would both change as the damper is inserted, this would be comput- edas:"4-4 ingustriat ventuauon side Plan View cross-sectional ‘ew of camper ET fare Elevation FIGURE 4-1. Blastgate damper (VP, +SP, ) (VP, +SP, Xenon = 8 Sue, , a, where: 1 = without damper inserted 2. = with dampor inserted some amount SP. = static presoure measured woll downstream of the damper mean velocity pressure a the seine location The more the required reduction in flow, the greater the value of Xaunper that is needed. Required values OF Xaenper om the same system for different branches typically range from to 5, Values less than 0.2 would have little effect on air- low, suggesting that the damper is not needed, For a duct that has a diameter that is much larger than needed, the value of ‘Xawuper could reach 100 or more. While i is clear that the greater the insertion depih of the damper the greater the value of Kamer, very little has been published! om the relationship of Xexmper insertion lepth ‘Crowder and Lowdermilk” published a table of insertion epths to achieve different pressies at a given velocity del’chik** provided a table of recommended values for a damper with a flat edge. Plotting them shows a highly nantin= ear relationship (see Figure 4-3), The resistance increases rmiich more than linearly, perhaps because: 1) a ane increas ingly inserts the damper, the unoccluded opening becomes progressively smaller, increasing the air velocity past the edge ‘ofthe slide, and 2) the region of separation becomes larger (see Figure 4-2), ‘As shown in Figure 4-3, the resistance due to a damper ‘changes litle until the damper is more than one-third elosed, vp = ‘Separation AAA FIGURE 4.2. Flow around a dampor At insertions greater than three-quarters of the diameter, large swings in airflow may occur with small additional changes in insertion depth, making adjustments frustrating and time-con- suming. Its much casir to balance a system ifthe balancing methodology minimizes necessary insertion depths so thatthe copstation is in a more level part ofthe curve in Figure 43. Figure 4-3 represents values for one damper. Other dampers may have different curves based on blade design and blastgate location, For example, the leading edge of the side (see Figure 4-4) can be straight or it can be rounded so that it follows the shape of the duct (convex) of rounded in the opposite direction (concave). Note that the concave shape precludes complete blockage of flow, a sometimes desirable trait. All dampers ‘would have relatively low Xaunper values when slightly inser- ed, but higher values as damper blades are inserted farther. As is known by analogy to orifice plates used for airflow ‘measurement, the sharper the leading edge of the damper the ‘greater the value of Xaampec at a given insertion depth, especial- ly ifthe beveled edge faces downstream. Since abrasive parti= cles in the air ean round the edge of the damper, itis possible 40 sai o 02 «04 «= 068s 1 Insertion Depth / Diameter FIGURE 4-3, Graph of Idelchik (1972) valuesflat convex concave FIGURE 4-4. Different leading edges that the resistance of the damper at a given insertion depth ean decline overtime with abrasive wear. Its also possible thatthe resistance would vary due to accumulation of contamination 6m the blade or just upstream or downstream of the blade. By creating a smoother path for the air around the damper, itis likely that an accumulation of contaminant in the duct would, sreatly reduce (not increase) the resistance, especially if it is fn the downstream side. On the other hand, if contaminant buildup reduced the unoccluded area, itis likely that resistance ‘would increase. ‘Once the required resistance of a given damper is known, it is also possible to use an orifice plate or other fixed obstruc~ tion with the same resistance in place of a conventional damper So-called “butterfly” dampers swivel about an axis to par- tally oF fully biock airflows, Ifthe blade is aligned to the duct, then the obstruction is minimal. If it is perpendicular then it will neatly or completely block the flow. These dampers are probably more likely 10 promote plugging than blastgates They cau be more difficult to adjust precisely than blastgates, especially when used to achieve substantial airflow reductions. For high velocities they may vibrate unless sturdy eomstract- ed, and they ean easily shift if they are not tightened very securely. 4.5 STRATEGIC GOALS IN ADJUSTING DAMPERS. Dampers and fans should be adjusted tozether to provide sufficient airflows to each hood while minimizing energy costs. It is prudent to assume that for a hood to be reliably effective its airflow should not fall helow some minimum reli- Balancing Duct Systems with Dampers 4-5 able value (Qgai) If Quo is Selected appropriately fora given hhood, then any amount of airflow above Qual is unnecessary and wasteful. Hence, for the example data listed as Table 4-1 (see Figure 4-5), Branches 2-A, 4-C, 5-C, and 6-D have exces- sive flows that should be reduced while the airflows in Branches 1-A and 3-B should be increased. The total airflow through the fan is 40% above the level that should exist ifthe system were perfectly balanced 4.5.4 Overall Goal. The appropriate overall goal in balanc: ing system airflows is to minimize the system airflow (i.e. the airflow at the fan) without allowing any hood airflow to become insufficient. This can be done by forcing the airflow (Q) in each branch to be as close as possible to its goal with- out falling below its Qoss. For the example data shown in ‘Table 4-1, the overall airflow before balancing was 40% above desired levels. a substantial excess. Mare importantly. same hhoods have less than their goal airflows. 4.5.2 Distribution Goal. The level of aicflow in cach branch is determined both by the fan and by the damper adjustments The purprse of tn daniis iow active dhe desiva Uisuibue tion of airflows (ie., each hood receives the desired fraction of the total airflow), The desired level of airflows can then be achieved by adjusting the fan output. Thus, the goal when «adjusting dannpers slivuld be wy free each bnaiile Uuctw vatty the same fraction (Qrais) Of its go as all other branch ducts. Tor a given branch: hate = (QW Qyoat) (43) For the example data shown in Table 4-1, values of Qrise before balancing vary from 0.82 to 2.57. Values of Qaio below 10 indicate insufficient airflow. Values above 1.0 indicate unnecessary airflow: 46.9.3 Tutal Atrftow Quak Fur ie syste ws. whole, Ue least energy is consumed if the fan airflow is the minimum possible value (Qpraa) and the fan pressure is the minimum that can exist when the fan airflow equals Qoos-an‘*!”” The Fiasco achieved whic all lauilies have aisflvws caacily eyual to their respective Qyat values. Since densities could vary throughout the systems, this minimum fan airflow can be com puted only from mass balance: TABLE 4+, First Example Problem D Dia Qneas | Qseai [ Grati +A 4 328 400 082 BA 5 529 400 132, 38 4 362 400 091 4c 5 | ssa 400 139 Be fe 5H 703 130 6D 7 7030 400 257, Total 357 2400 14046 Industrial Ventilation Plan View * Elevation View FIGURE 4-5. Example duct system atts Ona =( ge Hoe ws Where: if i branch duct 1 td meer ot enh co a = Gay euro oat Chinas = anew para If the static pressures upstream of the fan are less than 10 “wg and if the temperarures in each branch duct are within a 30 F range, the density factor is approximately the same throughout the system. Equation 4.3 can be simplified as: Gor ou = 2 Oar a1 FFor the example data shown in Table 4-1, the value of Qin ‘et IS 2400 acfin. 4.5.4 System Pressure Goal, The second goal in balancing should be to minimize the fan statie pressure. In addition to increasing energy costs, unnecessarily high static pressures stress the ducts and air cleaning device and may make damper adjustment more difficult. Fan motor loads are roughly pro- portional to both airflow and total pressure, Hence, an unnec- essarily high total pressure requires unnecessary energy expenses. Depending on the system resistance and the fan's operating curve, high pressures could also have substantial effects on the fan's efficiency and its noise level. A high static pressure also requires a higher rotation rate, possibly ata level beyond the current fan's upper limit. The minimum possible system pressure will occur if the system is successfully bal- anced while also leaving at least one damper completely open) As is discussed in the following sections, the interaction between system resistance, fan output and branch airflows complicates both adjusting the fan output and adjusting dampers 4.8 DETERMINING INITIAL CONDITIONS Some practitioners stat balancing with some dampers par- tially closed. If the airflow in each of those branches is just ‘below Qpoai oF is higher than Qyeas then opening all of those «dampers fully before beginning to adjust may be unnecessary. However, its probably safest to always determine initial air- flows (Qops’) with all dampers fully open. In those cases, i is necessary to protect the fan from electrical overload since the airflow and power requirement will increase when dampers are open. Iis also important to mesure hood statie pressure, SPs, and other key variables (e.g, fabric filter pressure drop). 4.6.1 Determining Airflow Goals for Each Hood. The air= flow required for each branch and hood (Qgeu!) should be no Jess than the values recommended in ventilation texts (eg.. Industrial Ventilation: A Manual of Recommended Practice Jor Design) In addition, the airflow should be high enough that the duct velocity exceeds minimum recommended values for adequate transport velocity. If the velocity is low, then either the duct must be replaced with a smaller one oF the Quo ‘value must be increased to Qgoai = Area"Vs 65) where: Vi = minimum velocity necessary to prevent setting A = Uross-sectonal area If a hood has substantially low airflow and is ineffective, its airflow should be increased to the recommended value, If a hood already has atleast the recommended airflow and its per- formance is poor, then it is possible that the recommended level is inadequate, However, one should first look to improv ing the hood design and the work practices since higher levels, of airflow with the same design and conditions may also prove ineffective If the hood design and work practices ure reasonably opti-‘mal, the airflow exceeds recommended levels, and the hood is ineffective, consider setting the value of Qua to be higher than the recommended level. It s important to not exaggerate the amount of increase. Ifthe level of airflow is already high, itis quite possible that increasing the airflows will not solve the problem, The next sections list three damper adjustment methods with a common example application for each. 46.2 Determining Adjustment Goals During Balancing. Although the goal in balancing is to provide the desired air- flow for each branch, when actually adjusting dampers, it may Ihe more convenient tose measited values that shold he prom Portional to airflow, such as velocity, or proportional to airflow squared, such asthe centerline velocity pressure or hood stat- ie pressure, Mean duct velocities can be determined from full 10 oF 20-point Pitot traverses of each branch (see Chapter 3). 46.3 Centerline Velocity Pressure Goals. Since velocity prsooure io proportional t the vollumotrie flow aquared (Q”),it is possible to use the mean velocity pressure as the indicator thet the airflow hae hoon adjucted 19 Qwar. Likewise, sinoe the ratio ofthe centerline velocity pressure (VP) to VParg should be reasonably conetant fora given meacuromant loestion, then ‘Pa also can be used as a surrogate for airflow. Thus, one ae oe , Q, or: WM Se = where: VPogesi = velocity pressure that should exist whon the ebeorved airflow aquale Agoat \Pacpen velocity pressure measured with all dampers open ‘The cemerline velocity pressure can be & poor surrogate for the volumetric flow if 1) its values are low (¢.g.. less than 0.15 “wy) due to low velovity ithe duct or to pressure sexis0% error, 2) the velocity profile varies duc to changing conditions upstream of the measurement, or 3) the velocity profile is highly asymmetrical. The latter is likely if VPjyy/VPa falls below 0.8 or execeds 1.0, If three measured values of VP at the center of the duct are greater than the other values itis likely that errors due to use of VP will not be excessive for purpos of adjustment when used in the manner described for this method. Its accuracy also is suspect if a single reading is taken, It should be measured three times, each time removing the Pitot tube from the duct and re-inserting it before taking the next runeagurernent, The median of the Uiree values should be used as VPs, Use of Equation 4.5 is not the same as employing the pipe factor (i.¢., VPuwp/VPa) with the assumption that it is always Balancing Duet Systems with Dampers 4-7 0.9, a common practice. Although 0.9 is reasonably accurate under ideal measurement conditions,*® under varying meas- ‘urement conditions pipe factor may deviate substantially from 0.9. Wang'** found in a study of five working duct systems that airflow estimates based on pipe factors done fora diverse range of conditions deviated by -25% to +10% from values based on 20-point Pitot traverses, Nearly half of all arflows that were estimated using a pipe factor of 0.9 deviated from traverse values by 5% OF more. For those reasons, 118 very important that VP. be measured at least SD downstream of any disturbance (¢ g,, elbow) and at least 3D upstream of any disturbance (especially the damper), 4.6.4 Hood Static Pressure Goals. Since hood static pres- sures (SPs) are proportional to velocity pressures for a broad range of airflows*"! a goal hood static pressure can also be ‘computed based on the SP, value prior to damper adjustments: ya Uyoas ¥ wan [te i = iP uate ieaidae exist at Q = Qgoa ea We ees ee Note: When measuring SP,, accuracy is improved by taking the median of three observations. Ifdensity does not vary significantly from the time prior to adjustments to the time the dampers are adjusted, Equation 4.7 simplifies to: Qa Shea = OF (=) a Since Que equals Q/Qeui, Equation 48 can be re-stated as SP ronal = SPrepe(Qrate® Hood static pressure values are a somewhat more reliable surrogate for volumetric low than are centerline velocity pres- sure values, but they, Wo, must 1) be high enough forthe pres sure sensor to measure accurately, and 2) be measured at least SD duwusieeaits antl AD upsticats uf distabauices, Like VFet values, itis prudent to measure SP} three times and use the yedian Valin fir ccgndatiat, 47 WHEN THE DISTRIBUTION SHOULD BE RE- BALANCED. If the airflows are all high in every branch, it may be only nevessary 0 adjust the fan airflow. Since every hood should48 Industrial Ventilation have sufficient airflow, the ideal fan speed would be such that the lowest value of Qui was 1.0. The desired airflow can also be obtained by adjusting the fan outlet damper if one is used. This achieves the goal of ensuring that every hood has enough airflow, but it does mot minimize the airflow require- ‘ment since some hoods may receive much more airflow than needed. Given the direct and indirect corte of balancing: good result ean be achieved if the final fan volume (Qa) oncoede Query by leer than 10% The excess system flow is affected by a variety of factors. It Is likely to be high if 1, The original airflow distibution was inappropriate. 2, The airflow requirement for one or more hoods changes. A branch is added or removed from a system, 4. A duet is replaced with one having a lager or smaller diameter or other changes are made that dramatically change resistance to flow: Owe vs tine Imad ave repetitions nt a anes that substantially changes the lengths and number of elbows in bianch ducts, or fit changes how and where the branch and submains are connected to each other. “Most systems may have to be balanced several times over their useful life, even with stable operating conditions and no major duct changes. When there are changes to operating con- ditions or major duct changes, systems are very likely to require re-balancing 47.4 Measures of Effectiveness of Damper Adjustments An obvious but misleading way to rate the effectiveness of ‘damper adjustments is to compute the relative difference between the sum of observed airflows and the sum of goal air- lows, Howould be micloading bevauce a ayctem with exoeccive airflows in some branches and counterbalancing insufficient airflows in others could have a zero “error” Indeed the total airflow could be made to match the total sum of goals by set- ting the fan speed without using dampers at all Clearly, any measure of effectiveness should consider the deviations of values ot Qos froma value of 1.0. For example, ‘one could employ the range of deviations from a value 1.0 in Example |. If that were done and values of Qnae ranged from (0.82 to 2.57, then the errors for individual branches would range from -18% to +157, One problem with that is that no hood should be allowed to receive insufficient airflow, If that ‘were remedied by increasing the fan speed to 1/0.82 times its previous value, the lowest value would be satisfied but the highest value would be 191% of design ‘The most usefil measure isthe relative difference hetween the lowest airflow at which the minimum Qacio i 1.0 and Qa = Qos geet Since the former is simply current fan airflow divided by the minimum value of Qa, this can be computed after adjusting dampers from: Qaienes = [Amin Qratc) * QtalQtangoat 11" 100% (410) where: min(Qrate) = minimum ratio of actual to desired airfiow among all branches (Qian = airflow at the fan determined at any stage. In the previous example, Qauy Qasr gaat ~ 1:40 end min(Quate) is 0,82. The total Quicas computed using Equation 4.10 is (0.82414 1)*100% — 71%, a lange amount of wasted air flow, 48 BASIC AIR BALANCE METHOD. ‘The Basic method is the most commonly used balancing procedure. Different individuals may have somewhat different strategies bu the basic idea isto adjust the first damper so that for it Q= Quo, then adjust the second damper so that for it Q = Qeoa and So on until all dampers have been adjusted 10 achieve the desired airflows. However, since VPa! is propor- tional to airflow squared, one ean use VP as.a surrogate for Q during adjustments. This method is the least aceurate of those discussed in this chapter, but if the procedure listed below is followed, each hood should receive airflow greater than or ‘equal to Qys 4.8.1 Procedure for Basic Method. The following isa ver- sion of the Basic balancing procedure: 1. Determine the desired airflow (Qso) for each hood. ‘Compute Qiangou from Equation 43 or 4.4. 2. Open all dampers, taking care to protect the fan motor since the power it requires may increase suhstantially Optional: partially close dampers for ducts whose ait- flys are known in he highly excessive 3. Conduct Pitot traverses and measure VP. and SPs for cach branch duet, compute the “open” damper value of Airflow (Qope) for each branch, 4, Compute the initial Qianopea fom the sum of the observed airflows. ‘5. Ifnecessary, adjust the fan's airflow until Qen ops! Qn gut exceeds of is equal to 1.15 and ie below 1.40. A value of 1.2 is probably the most prudent adjustment goal if that is dono. 6. Compute VPz.gsat for each branch duct using Equation 46. 7. Determine the order to adiust dampers based on con- venience, decreasing duct size, or increasing insuffi- ciency of airflow. 8. Adjust each damper in turn unti itis observed that VP equals VP tgs! for that branch, Note that only the last damper adjusted will have the desired airflow. All oth- crs will change as each damper is moved.©. Repent Stp 8 (in, 2 round). 10, Ifnecessary, repeat Step 8 again (ie, 3 round) 11, Do a fl Pitt traverse for each branch duct deter rine the final observed airflows (Q) and all Quo val aes (Equation 4.2, 12, Ifthe lowest value of Qn min( Qs) less than 1.0, increase the fan spocd based on: o, = 2 min Qeae Noe tae ey Harp ser ie a alow ta above the minimum level as a safety factor, 1f the inital total aielow is adequately high but not exces- sive (e8. 115% to 150% of Qin this procedure should ensure that every hood has enough airflow and produce lose tothe desire dstnbution into or thre rounds of adjustment Using this method, Halasubramian>' found the excess alow volume to be approximately 4.8% to 8.5% in two rounds of adjustments ofa system with seven branches. For these tess, the measurement conditions were ideal and initial airflows ‘wore sufficient, The test started with all dampers fully open. then adjusted dampers in the order of most excessive to least excessive Quuo values. Under field conditions the excess ir- flow likely to be somewhat higher in many cases 4.8.2 Example 1 (Basic Method). The example application listed ere is intended to illustrate the use ofthe Basie method, ‘There are other methods that wil also show relative advan- tages and disadvantages, Follossng the steps ofthe Basie Method, the results would be as follows 1. Determine the desired airflow (gad for each hood. Compute Qtan-goa from Equation 14 Note, if the density changes within the system, use Equation 4.3. The values shown in Table 4-2 were selected abirrily for this example TABLE 42 intial Measurements ana Atiow Goals tor Example Problem Branch | Dia_| Qopen [ Gyo | Gato | SPh TA 4_[ 326 [400 | osz | 110 Ezy a 38 4_[ 32 [400 |_omt_| 1a a0 a a 3e =| 3 | 400 | 139 | 166 co 7_[ 100 [400 | 2a7_[ 182 Fan [14 | sa57_| 2000_[ 140 Fan soeed = 696 rom Balancing Duct Systems with Dampers 4-9 Open all dampers: Optional: pertially close dampers for ducts whose airflows are known to be highly exces- In this example, all dampers were opened fully Do Pitot traverses and measure VP.) and SP for each ‘branch duct, then compute the “open damper value of ‘airflow (Qopen) for each branch. ‘The results are shown in Table 4.2, Quegua = 2400 acti. Compute the initial Qjeeopen from the stim of the ‘observed airflow From Table 4-2, Qraropen= 3357 acfim for Qanapen Ynocaesary, adjust tho fam’ airflow stil Qn Ojo gost exceeds or is equal to 1.15 and is below 1.40. A Walne of 1.2 is probably the mast prudent adjusoment goal if that is done In this example Qtar-ope/ Qgoatian = 3357 act / 2400 acim = 1.40 Hence, the fan output was not adjusted Compute VPetgoi for each branch duct using Equation 46 ‘Table 4-3 shows the measured values for this example For Branch I-A, the measured VP wos 0.98 “We, Since Qrio WAS 0.82, VPs gu) = 0.98(0.82) = 1.46 owe. ‘The remaining values are shown in Table 4-3 Determine the order to adjust dampers based on con- venience, decreasing duct size. or increasing insuffi ciency of airflow: The order is shown on Table 4-3. Adjust each damper in turn until tis observed that VP.) equals Vigo for that branch. Note that only the last damper adjusted will have the desired airflow: Alt oth- ‘ers will change as each damper is moved. For a given damper, increase and decrease the damper insertion depth until the observed VPui is as close as possible to VPyisoa. AS can be seen in Table 43, Branch 1-A was deficient in airflow, so its damper was not adjusted in Round 1. Note also, VPui was less than \VPcrgt. Since the damper was full open, any change in the damper position would decrease VPg, a worse condition. Next, Branch 2-A was adjusted until it was 401 actin, (very close to the goal of 400 actin). After that adjustment, all other branch airflows would have ‘increased but the airflow through the fan would have fallen slightly. measured, iis probable that Branch 3- B had risen above 400 acfin.410 TABLE 4.3, Basie Method - Intal Round Measured ‘Goal Round Branch Qiao Va Vege Order Resulting @ Gate 1A 0a 088 188 7 2a 731 2a 132 105 060 2 572 138 8 ost 120 188 3 882 138 4c 139) 118 060 4 540 138 5S 139 116 060 5 500 125 3 257 108 078 é a0 101 Total 140 3082, 129 ‘TABLE 44, Basie Method ~ Second and Third Rounds 2 Round 3 Round © Order Resulting @ rato Order Resuiting @ ‘ratio TA 1 480 120 1 42 103 2A 2 8 112 Zi 08 1.02 38 3 6 1.09 a aot 107 + 4 20 1.05 4 40 101 56 5 8 102 5 404 101 6D 6 300 100 6 400 +00 Total 2692 1108 2432, 101 If Branch 3-B was now higher than originally but still bolow 400 acfin (or at least VP. was less than VP soa), We Would leave its damper alone. Otherwise, we ‘would reduce the damper opening until VP equals VPs If we now measured the airflows in Branches L-A,2-A and 3-B, it would have indicated that Branch, 1-A was higher than originally measured but till below 400 actin, Branch 2- had risen even further above 400 acfin, and Branch 3-B was somewhat above 400 acfin, Adjusting Branch 4-C set its airflow temporarily at 400 cfm but increased airflow through the other branches. In particular, the airflow in the branches previously adjusted t0 400 acfin will increase above 400 acfin. A similar result occurs when adjusting Branch 5-C Adjusting Branch 6-D required inserting its damper ‘more than 90% into the duct, Much of the airflow from Branch 6 was diverted to other branches, producing the results secn in Table 4-3 at the end of Round 1. The col- umn “Resulting Q” contains the airflow values after the first round of damper adjustments. Note that the total airflow is now 29% abave the sum of Quest Values and 10, that the Quis forall ut Branch 6-D was atleast 25% above goal levels, A 29% excess volume would indi cate another round should be attempted, ‘There isno point in doing Pitot traverses yet, but if they were done the results would be the values of Quai shown on Table 4.8 Repeat Sep 8 ie, 2 round) ‘Complete the second round of adjustments, As in the first round, each branch airflow is adjusted in urn until st reaches pal, but as each damper is adjusted the air ‘lows in previously adjusted branches rise increasingly above Qpnu levels except the last damper adjusted, Branch 6-D. As shown in Table 4-4, Branch 1-A has a 20% excess in flow but the fan airflow is now only 8% cover goal values Since the ~worst” branch is 20% above the goa, a third round was begun, necessary, repeat Step 8 again (Le. 3* round) As shown in Table 4-4 atthe end of the third round,all branch airflows were within about 3% of goal lev els and the overall excess is roughly 1%, No more rounds were required, 11. Conduct a full Pitot traverse for each branch duct to determine the fino abeored mirflowe (Q) amd al3 Qos values (Equation 4.1). ‘The resulting Qace Values are shown der Round 3 on Table 4-4. 12, Ifthe lowest value of Qrats(in(Qna)) is less than 1.0, Increase the fan speed based on a min Orato oO In this example the lowest Quo had a value of 1.0, hich should always be the case ifthe initial fan speed ‘exceeded ideal levels. No fan adjustment is needed. Notes Alter the third round of adjustments, the error was about 1%, well within normal expectations. Although a good balance vwas achieved after three rounds of adjustments, this ean be very time-consuming. To achieve the required resistance through the dampers, especially on branch 6-E, the damper ‘must be nearly completely closed. In that ease, even the slight- est change in insertion depth would have large effects on ai flow, making it very difficult to set the damper so that the a flow equaled Qgoa. Finally, a by-product of such high resi ances isa high estimated fan total pressure of 14.4 “wa, a rel atively high value Some practitioners have used the short cut of taking a cen- terline reading for VP and then applying a factor to estimate velocity (and volume) ina duct. As discussed earlier, the errors from that estimate can be quite large. Iis important to take fall Pitot traverses after balancing to document the final results Ina systern that was originally balanced for which re-bal- ancing is required, it may be reasonable t0 start with the dampers at the postion found, rather than beginning by open- ing all the dampers. The effort to balance the system may be considerably less. However, this approach may net optimize the performance of the balanced system. The additional fort ‘uTbalancing assuviated wid opengl Uanmpeis i Ufc jus tied. 4.8.3 Advantages and Disadvantages of Basic Method. The procedure’s advantages are that itis easy to understand and it requires litle preparation. Its disadvantages are: 1. The procedure will fail ifthe fan output before balane~ ing is not sufficient. Otherwise, as the damper blades are inserted and the fan airflow declines, there is not enough airflow for the last branches adjusted. If this ‘occurs, one must inerease the fan speed and hegin again, Duet Systems 2 Conversely, ifthe total airflow is initially significantly higher (e., more than $04) than design, the damper adjustments will be more difficult because the dampers pethaps tothe point where even small changes in insertion depths have substantial effects on flow (see Figure 43). 2 Avaach damper’ insertion depth i changed ce thatthe current branch has the desired airllow, airflows in all ther ranches are inceeasing from the originally adjusted levels. In addition, the fan airflow declines \with each damper blade that is inserted, possibly to lew els that are insufficient for the system or that produce velocities in some submmains and mains that are too low: 4, Ibis time-consuming, To achieve acceptable arflows, it Js almost always necessary ro conduct two and some- times three rounds of adjustments, For each round, ai ows may have to be calculated several times, 5. Because the dampers are being used to “choke down’ the fan arflows in most cases, the fan pressures and fan ‘motor operating costs are typically higher than could hhave been achieved with methods that separate achiev- ing the desired relative airflows from controlling the fan output. 49 INTERACTIONS BETWEEN THE SYSTEM AND ‘THE FAN ‘The airflow output of a fan varies inversely with resistance flow! A danas aid sistant hn ut in white iti installed and thus to the entire system, As dampers are insert- al ands fl a duns che fn as flow (Quan) alsu Uerisasen. At any change in system resisince, the change in fan airflow \Joyinds vn the slope uf de Fa’ aifhow-pressuic Lure at dia level of system resistance. The efficiency of a fan also varies, with resistance of the system and therefore with changes to damper insertion depts. Some air cleaning devices (eg, fabric filters) also are vari- able sources of resistance. They complicate balancing because they produce changes in system pressure and airflows inde- pendently of the damper. 4.9.1 Throttling the Fon Airflow. The resistance aided by « damper not only changes the nroperton af the fun selene ‘going to that branch, but also the total airflow delivered by the ff (Gee Figure 4.6) This interaction complicates the adjnst- ‘ment process, It also makes it possible to reduce the fan air- flow toa desired total oumpit sing the dampers while alsa achieving the desired distribution Although fan inlet or outlet dampeis can be utilized 1 reduce the total fan airflow to desired levels, there are disad- vanlages for this method when vonpared wo eluate Can speed: 1) Higher fan pressures result in resistance that is higher than necessary for distribution purposes, increasing412 Industrial Ventilation ‘System with dampers ; System without dampers TPran / Fan Curve FanQ, FanQ, FIGURE 4-6. Effect of dampers on fan airflows operating costs and fan speed. Fan speed inereases can possibly exceed the safe maximum for the fan, 2) Fan efficiency may be affected adversely ifthe fan was initially selected without allowing for the additional resistance of the system due to the dampers, increasing ‘operating costs, 3) Dampers must be inserted farther, possibly making adjustment more difficult and inereasing the chances for material build-up. Inctoad of “throttling” down the fan withthe damper, itis possible to separate damper adjustment and fan airflow regu- lution. This is dono by adjusting branch duct dampers to achieve the desired distibution (ie, fraction) of airlows instsad ofthe desired love of aicflows, Once that distibution is achieved, then adjust the fan srflows (ideally by changing the rotation rate) to obtain the desired tal fan output. Since the distribution of airflows is independent of the level of air- flowa(#% changing the fan speed doco not ordinarily affect the cistribution 492 Optimum Initial Fan Output. Kan cuttlows wil decrease as dampers are inserted. It would be very helpful to be able to predict the airflow that the fan would produce after all dampers had been adjusted. Ifthe Basic or Modified Basic Method is used, it would allow setting the fan rotation rate to an optimum value before adjusting dampers. For the Target Method, discussed in a later section, the prediction would allow more accurate determination of target parameters (€.., Pp ot VPs). may be possible to estimate the final fan airflow (Qin is that would exist ifthe dampers were adjusted to produce a per- fect relative distribution at the lowest possible damper inser- tions (Quai). The use of this predicted value would depend on the halancing strategy. 4.9.3 Estimating Qranigear Without sophisticated models, one cannot reliably predict Qin-imi, but it may be possible 10 estimate it reasonably well. For example, multiplying Qiargoa, by 1.2 (a 20% increase) provides a rough estimate. This sim- ple estimate will probably be sufficient for most systems that, are being re-balanced. For example, if the value of Qimon iS, 3300 aofim and the value of Qi 8 3000 acti, then the esti- ‘mated value of Qe sat Would be 3600 acfin. The fan output, should be increased to 3600 acfin before adjusting any dampers, Howeves, the ideal level of Qin may he much less than 1.2*Qnn ga. For example, ithe relative airflow distribution is already very close to ideal and Que apn is neuly the same as ‘Quango then Qian ssa Should be only slightly greater than Qu Temay be possible to estimate the final fan airflow (Qn ns) ore closely of 10 at least put boundaries on the possible range Of Qisnéaa Values. The lowest possible value would occur vwnen the Tan aitiow 1s mdependent of damper resistance. in that ease, the effect of inserting a damper is solely to shift air- flows to other branches. The final value of Qén Would equal Qian pen ad all Quis Values would Be equal {0 Qra-ope/ Qen The other extreme occurs if the fan airflow is so affected by damper adjustments thatthe fan airflow falls exactly by the amount the airflow to a branch is reduced by adjusting its, damper. ln that case, the value Of Qiao forall dampers must ‘equal the lowest intial airlow ratio, min(Qasn) and the final vale of Oss must equal the Oso divided by the min( Onis) The value of Qi. it could fall anywhere betwoen Qierasn and Quangon ‘min(Qrao). Where it fall in that range would depend on the values of Néanper for all dampers as well as the fan curve. However, a value halfway between the fro values, Qn sccat = (reper + Qrangoatémin( Grae V2 ‘may be a conservative estimate. A better, although more ‘complicated estimate is given in Equation 4.11 arte = (Cher cpon)/ 4°{Qhar open goat + min(Qret)} [AN] I Equation 4.11 were applied to the data in Table 4-2, the proto iain initial aislow slvuld be Qten snes = 9957 / (0.5)"(8957/2400 + 0.82) = 3957/1.11 = 3026 acfm ‘Since the imtial measured value was 335/ actm m this example, the ideal value is about 10% lower than the initial, value, 1f one assumes that the tan output wall be adjusted after adjusting all dampers.A study of typical systems and fans selected for them would likely find a mean optimum value that is somewhat higher or lower, but the midpoint can serve as a reasonable estimate until, ‘a better estimate is determined 4.9:4 Fan Adjustment After Balancing. cr all dampers have been adjusted, all branch airflows should be measured sand npc Qppt vals, Sins primary gl i or cach hood Q to match or exeved its Quo valu, all values of Qaso should match or exceed 1.0. The fan spcod should be adjusted from its current rotation rate (co) so that the branch with the ‘inimum Qaio value has sufficient airflow. That will be accomplished ata new rotation rate (0) that can be computed fu ©. o, =o 412) mir(Qraio) where: min(Qhais) = minimum Osis value after all dampers have ben adjusted For example, suppose that after all dampers have been adjusted, the values in Table 4-2 were measured. The total air- flow through the fin is 40% above the minimum desirable amount, but the airflow through two branches are below 100% of their goals. To make sure that Branch 1-A has sufficient air- flow, it would be necessary (0 inerease the fan rotation rate using Equation 4,12; 62 01 / Min(Qrato) = 01 | 0.82 = 1.22 4.10 MODIFIED BASIC METHOD The moat severe diffisultica inthe basie procedure gencral ly are due to extremely high or low fan speeds and problems ‘vith estimating airflows using centerline velocity or velocity pressure readings. The Ventilation Manual strongly recom- mends against using the centerline velocity, Vain calculating the average velocities and volumetric flow rate. However, the problems that oveur from this ean be addressed in this seetion ina modified version of the Basic procedure. Before adjusting the first damper, the fan speed is adjusted to achieve sufficient fan airflow, Altematively, the target airflow can be adjusted to aie the witflow available wx che wounyption dist dhe fun speed will be adjusted after balancing the dampers. ‘To reduce the errors associated with using VPa, dampers are adjusted using measured hood static pressure values (SP,) rather than VP. SPs has been shown to be highly correlated to airflow squared” Another difference is that the fan airflow is always adjusted to the “ideal” value prior to adjusting. any dampers. 4.10.1 Procedure for Modified Basic Method. The steps for the Modified Basie Method are: Balancing Duet Systems with Dampers 4-13 1, Determine the desired airflow (Quy) for each hood Compute Qin sot fom Equation 4.3 or 4.4 pen all dampers, taking care to protect the fan motor since the power it uses may increase substantially Optional: partially close dampers for ducts whose ai flows are known to be highly excessive 3. Do Pitot traverses and measure the volumetric flow, Q, and static pressure, SP, for each branch duct, then compute the “open” damper value of airflow (Qspe) for cach branch. Measure and record the amperage for the fan motor, the SP across the air cleaner (AP), and ‘other values you may wish to document or use later. 4. Compute the init ‘absorved airflaws, I Qtansper from the sum of the 5. Determine the ideal fan airflow (Qtaiga). This ean he 1.2* Qlingol the average of the Qfan goat 3d Qin opens br the value determined by Bquation 4.11 6. Adjust the fan speed from the current value, 9, too 2 — 04 X Ordeal Arras 7. After adjusting the fan speed, measure SP}, for cach Lbanch duct 8. Compute the desired final hood statie pressure, SP gel for each branch from Equations 4.7 or 4.8. Compute the values of SPy/SPh.gai using the values of SP, meas- ured atier adjusting the fan speed. 9 Beginning with the largest ducts and continuing through 1 the smallest or with the highest value of SP\/SPingai and ending with the lowers, adjurt each damper in tur until the measured SPh equals SP} for that branch, Optional: adjust the first few’ dampers to achieve SPi values that are 3-10% below the SPiygcl_ Values Adjusting the hood static pressures slightly lower than the goal compensates for the general aurtlow merease as subsequent dampers are closed 10, Repeat Step 9 fora second round 11, Repeat Step 9 fora third round, if noeessary. 12, Do a Pitot traverse for each branch duct to determine the final observed airflows (Q) and Qevio values. 13 the lowest vatne of Qnsie, min(Qras) lass than 10 imrease the fan speed based on: Note that it may be prudent to sct the fan airflow to 5% above the minimum level (2 ) asa safety factor 4.10.2 Example 2 (Modified Basic Method). This example will once again use the data in Table 4-2, which represents the system in Figure 4-5. In the Modified Basic Method, one414 always adjusts the fan output 10 Qaida unless itis already clase (eg, no more than 5% in excess OF Qfaniea- That step should assure that airflow is adequate without being excessive (Steps 1 thru 4 are the same as Basie Method Example) 1, Determine the desired atrlow (Ogsx) for each hood. Compute Oyo from Equation 4.3 0F 4.4 ‘The values shown in Table 4-2 were selected arbitrati- 1y for this example. Open all dampers. Optional: partially close dampers for dicts whos aicflons re Inown te be hy axcee 1 this case, all dampers were opened full. Do Pitot traverses and measure Q and SPs for each branch duct, then compute the “open” damper value of airflow (Quer) for each branch ‘The results are shown in Table 4-2. 4. Compute the initia Qs from the sum of the observed airflows. In this example, Qunepen = 3357 actin and Qar-gou!= 2400 actin. Determine the ideal fan airflow (Qfmate {is can be 1.2 * Urn goa the average OF the Urn gst and Qsncpor oF the Value determined by Equation 4,11 In this example set Quovidea = 1.2 * Quogost Qtamidel = 1.242400 = 2880 acfim, Different estimates of Qfarial could be used. Some provide less adjustment as the method is used; however, they are all estimates of the idcal volumetric flow. 6 Adjust the fan speed from the current value, 1, t0 0: ‘Noting that the original rotation rate was 598 rpm and. that Qrin-open Was 3357 cim, Equation 4.12 can be used to determine the “ideal” new fan rotation rate: ‘TABLE 45, Modified Method Example ~ Initial Adjustment crt X Qtanssea! Qan-opan = 598 rpm x 2880 acfm/3357 acim = 513 rpm 7. After adjusting the fan speed, measure SP, for cach branch duet After adjusting the roration rate to $13 mpm, the new values of SPs that should exist are shown in Table 4-5 under the column headed “Measured after fan speed adjustment.” 8. Compute the desired final hood static pressure, SP). sear flor each branch from Equation 4.7 or 4.8. Compute the values of SP,/ SP.eua using the values of SP} measured after adjusting the fan speed. ‘The resulting values are shown in Table 4-4 under col- ‘umn heading “Compute SPingst 9. Beginning with the largest ducts and continuing through 10 the smallest or with the highest value of SPWSPh.gat and ending with the lowest, adjust each damper in turn until the measured SPs equals SPs goa! for that branch. ‘Table 4-6 shows the measured results after the first ‘damper adjustment. The calculated column Quint = Qi Qoat shows a roacure of how clove the firt round of adjustments io to achieving the goal. Since the ratio for 6-D is 1.21 additional adjustment chould be made. In thio oxample all the branches are adjusted. In a real world situation, only those that are above 1.05 or possibly 1.1 would be adjusted. 10. Kepeat Step 9 for a second round. Adiust the dampers a second time, After adiustmen ‘measure the volumetric flow and the static pressure in each branch, The results for the second round are shown in Table 4-6 under the column hea “Measured second round.” Goal Measured Calculated Measured after fan adjustment ‘Branch peat Cis ‘Phonan Oo Shgen | Grteton SP 1A 200 38 78 082 175 2 oar 2A 400 529 151 1.32 086 454 1m 38 0 362 146 oat 176 ant 7.08 + 400 384 68 139 087 475 122 5G a0 04 788 1 oer a7 12 =o “00 “020 198 268 020 a vat Fan 2400 3357 2000TTARIF 4. Madiiad Mothod Feampla - Fret and Second Rounds Duct Systems with Dampers 4-18 ‘Measured first round Calculated Measured second round Calculated Branch On SPha Qraves Oya SPuz Qiao +A 407 181 1.02 402 178 tot ZA ‘a0 oy 102 405 os Tot aa ‘ann TR 109 ane 17a 10 4c 432 101 108 420 0.95 105 5c 465 17 116 410 ost 108 60 483 02 121 18 032 104 Fan 72009 200 LL, Repeat Step 9 for a third round, if necessary. As shown in Table 4.6, the values of Quio2 (= Qi 2/Qgpai) were all below 1.05. Consequently, it was not necescary to have a third round of adjustments 12, Do a Pitot traverse for each branch duct to determine the final observed atrffows (Q) and Ono values The results are shown in Table 4-6 (Qu2 and Quai 13. Ifthe lowest value of Qnoi min(Qnaioh is less than 1.0, inorease the fan spaced haved on: f_o _) (areata) Since the minimum Quio Was 1.01, it was not neves- say 1 lange Ue Gar speed. However, © dennustiate the approach, Table 4-7 shows the results when the fan iy slowed down by 1%, Cu reality, a cluange Fess dha 5 percent would not be made.) After adjusting the fan speed, die results ate shuw in Table 4-7 ‘Notes on the procedures: 1) Although the procedure determines the required fan speed, st is nen not practical te achieve exactly the desired fan speed. The sheaves (pulleys) are limited and only 9 apache ce br pcterrued Nicol ‘practice is to run the fan faster thatthe calculated value, This is less important on the first fin speed! adj since a final speed will be required. The extra fan speed is essentially a safety factor Care is required to ensure that safety factors do not compound resulting is an excessive exhaust rate 2) Some designers adjust the final fan speed to a nominal ‘5% above that desired for an additional safety factor. 3) In the above example, the volumetric flow after each adjustment decreased from 3024 aefim after the first fat speed adjustment, to 2474 actin after the first damper adjustment, to 2393 aefm after the second fan adjust. ment, This progression is typical with each damper adjustment as the dampers are inserted farther and the system resistance increases. The fan backs off on the fan curve. 4) ‘The final speed increase was not to increase the flow through the fan to the goal, but rather to increase the flow through each branch to meet or exceed the goal for the branch, 5) The efficiency of the balancing is shown in Qe. If Qescess is 100 lange, it may be economical to continue the balancing, However, a single branch that signitt- cantly exceeds the goal may not warrant the rebalanc~ ing effort unless Qu is also excessive, 4.10.3 Using Adjustment Targets That Are Not Equal to the Goal Values. In many cases the target airflow (Q) during balancing should differ from Qe! values for reasons discussed in this section. Balancing systems is complicated by the fact that airflows can change in a duct even when its damper is not touched. This ean happen because: 1) Qian fluctuations due to the air-cleaner (AP changes as bags become dirty, etc.), 2) changes in resistance of branches due to extemal causes, 3) TABLE 4-7. Modified Method Exemole - Final Conditions Modify fan Branch Q Error A 400 0.0% 2A 403, 0.8% 38 402 04% a a Lae 5c 408, 2.0% 6D a5 35% Fan 2446 19%4-10 Anuustrial Yenutadon changes in Qin due to adjusting other darapers, and 4) shifting ‘of flows among branches due to other dampers being adjusted For example, suppose thatthe pressure across a fabric filter varies over its cleaning cycle from 2 10 S “wg, producing @ 10% change in fan airflow over the eyele. During damper adjustment, the airflow achieved would depend not only on the ‘damper but where the fan was inthe cleaning eycle. The same ‘would be tre if branch resistances changed for reasons having nothing to do with the dampers. For example, a floxible duct connected to a moving hood would experience changes in resistance, Ke share of airflow would fall whon its resistance increased and climb when the resistance decreased. All other airflows would change in the apposite dirsction in each ease A damper set to achieve Q = Quai may only achieve it for one shape ofthe flexible duct In both ofthe cases above, it would be difficult to compen- sate exacly for these events. The best strategy may be simply to adjust toa target tha is slightly greater (e, 5%) than Qua Some airflow will be wasted some of the time, but hood air- flows would seldom be insufficient 4.10.4 Branch Airflow Changes. Changes in airflow through a branch due to other interactions can be both areater {neffectand easier to compensate for accurately. The two most important interactions are those: 1) between branches and 2) between branches as a group and the fan. If one adjusts a damper to achieve a particular airflow in a jiven branch, the actual airflow will rise as each succeeding branch's damper is adjusted since their relative resistances are inereasing. If one adjusts each damper to achieve Qu, only the last branch adjusted will have Q = Oya. For all others Q > Oyu. Hence, the target during adjustments should be lower than the goal values forall except the last branch adjusted, and the branch- es adjusted first should have the greatest reduction, A tril and error experiment produced the following relationship For velocities or aitflows: k = (n/NPO445 (414) where: n= rank order number of the branch N_ = total number of branch ducts, k= order factor Table 4-8 show approximate k values for ranges of the value of nN. Another correction can be made for insufficient or exces- sive fan airows. 17 Quest Is defined asthe value of Qin that ‘would exist if dampers were adjusted perfectly and with the ‘minimum system pressure, then the diference berween Qu onen and Que epresents wasted airflow, Instead of reducing the fan speed before adjusting dampers, one could simply increase target airflow, Q,, proportionately to the excess. This ‘Would allow damper adjustment to acmieve a relative disbue tion which would then be followed by a change to fan airflow. For example, if Qprcpes/Qa.isat ~ 1.5, one could set Qr = 1.5 x Ques for each branch, The fan speed could be adjusted after balancing instead of before. Hence, considering both the fan and branch-to-branch interactions, Que Could be comput ed from: KY Qa! Oar F INO, omw-(as | 2 a a) Likewise, the target hood static pressure and the target cen- terline velocity can be given by Qn ee /Qan os + MINOW ous) | Pa Methods hased on hood static pressure measures (SPs) are ‘more robust than those relying on centerline velocity pressures (VP) because VP, measurement accuracy is strongly affec ed by disturbances. In addition, the ratio of VPa to Q? can vary semawhat fini ane meaciremant ta anather SP, ie alea ma cr to measure than VP, since itis not so critical in measuring SSP, vale that the penhe he in the exact center af the dict er that the probe be aligned perfectly with the airflow. Locations for SPs, measurements also typically are more easily accessible than good locations for VP. or for Pitot traverses, VPu should be used instead af SP, only if measiirement af SP is infeasie ble. 1c is recommended that dampers be located three or more ‘duct diameter lengths (3D) downstream of the SP, measur ‘ment location. 4.11 TARGET METHOD The third method is somewhat more complex mathemati- cally than the second method but it reduces the measurement efforts. The method reauires the same initial measurements. ‘The method is the same as the Modified Basic Method through Step 4. Like that method. this one uses SP measurements to adjust dampers. Unlike the Modified Basic Method, the fan speed need not be re-set before adjusting dampers, TABLE 4-8. Approximate k Values for Damper Adjustment uN k nN k 0010 080) 0510.60 030 011-020 093) 061070 0.98 021-0.30 085) 07080 098 UsToa0 cy ey 10 047-0.50 287) 0944.00 1.00)4.114 Procedure for Target Method The steps for the ‘Target Method are: 1. Determine dhe desired flow (Qed) for each hood ‘Compute Qing from Equation 4.3 or 44, 2 Openall dampers. Optional: partially lose dampers for ducts whose airflows are known to be highly excessive In this case, all dampers were opened fll: 3. Dio Pitot raverses and masite VP. and SP fir each branch duct, then compute the “open” damper value of Airflow (Qapan) for each branch 4. Compute the initial Qowapen from the sum of the observed inflows 5. Do NOT aust the fan speed prior to balancing unless Qlnopen is less than 0.7 oF greater than 1.5 times Qa soul HF 50, compute Qen.it from Equation 4.11 and adjust the fan output to achieve that value, then re- measure values of SP». 6. Rank order and number the branches hased on their mis values from 1 10N Ascion orter factor (value for each branch duct based on its rank order number using either Faqatinn 4 14 or Tahle 4-7 7. Compute the desired final hood static pressure, SP fiom Equation 4.8 and the wgst hood staie pressure (SPitnze) for cach branch from Equation 4.16, 8. Beginning with the branch with the lowest value of Qa and contimiing through to the next to smallest, adjust each damper in tum until the measured SPs ‘equals SPh.sr for that branch. Leave the branch with n= N completely open forall steps 9 After the Fret rnd ic complete mensite SP, for each branch again 10, Compute SFkpe ~ SFWSPsst for Gaets brats an {determine the median value ofthe ratios, median (SP). stn. 11, Adjust cach branch damper so that its measured hood static pressure equals: SPu= SPhgut x med(SPL su) Begin with the duct whose SP value is the great- cst, followed by the Jeast and altesnate between next highest and next lowest until roughly one-half of ciampets have been adjusted a second time, If neces sary adjust all dampers. 12, Do a Pitot traverse foreach branch duct to determine the final observed arflows and Quy Valves 13, Using the minimam Qeuo value determined above (rin(Qais)), adjust the fan spced from the original rotation rate (01) 0: 02 e/min(Qoaw) ‘ote: Some designers set the fan airflow to 5% above the rinimum level (o:) asa safety factor. Care should be used to Balancing Duct System svoic componnding saery factors 4.11.2 Example 3 (Target Method). The same example use for the Basie Method and Modified Basic Method is used haere to demonstrate the use of the procedure. Given the initial zasurermients and airflow goals Histed in Table 4-2 and Figure 4-5, follow the procedure listed above to balance the dampers The step by step solution is as follows: 6 Dotormine the desired airflow (Qn) for 200k hood Compute ang from Equation 4.3 07 44, ‘The values shown in Table 4-2 were selected arbitrari= 1y for this example. Open all dampers Opvional: Panially close dampers tor duets whose sit- flows are known to he highly excessive kn this case, all dampers were opened Filly Do Pitot traverses and measure VPui and SP, for each branch duc, dhe empute de“ open” darper value oF sirflow (Qopen) foreach branch, ‘The results are shown in Table 4-8. ‘Cecile Ws Sate Ragan ie EN observed airflow. Fron Table 42, for Qinopsn ~ 3957 ali Do NOT aust the fan speed prior to balancing unless kin open is Ie than 0.7 of greater than 1.5, To this example the ratio was 1.40, se adjustment prior fo balancing is necessary. Rank order and number the braucties based on Ur Qnso values from | 10 N. Assign order factor (k) value for each bail: duet based Um its rank order number using either Equation 4.14 of Table 4-8, Compute the desired final hood static pressure, SPp. sot and the target hood static pressure (SP) for each branch from Equation 4.8 and 4.15, respectively ‘The results are shown under the column labeled “Compute Initial Target. For Branch 2-A: SPhgoat = SPropen / Orato? = 1.91 1 1.92? = 0.87 “wy SP hao =| (5 | See 0 | Ph / ye 98 )/335%ef . | au) samo 22)) 51 = Beginning vith the branch with the lowest value of Quai and continuing through to the next smallest, ‘adjust each damper in turn ntl the measured SP, equals SPh.onst for that branch. Leave the branch with n= ¥¥ completely open for all steps.418 Industrial Ventilation Aer the frst round is complete, measure SPs for each branch again “Te raul are chown in Table 4 9. Note tha the route ae shown under the column labeled * Measured after first round.” Note thot the vahuoo ars much larger than the goal values 10, Compute SPinain= SPi/ SP OF €ach branch and determine the median value of the ratios, median (SP, at. Note that the ratios of the hood static pressure to medi an range from 0.83 t0 1.08 Adjust cach branch damper so that its measured hood atic pressure equals: SPhinnot = SPrgoa” Median|SPh to) Beggin with the duct whase SPh.nin value isthe great- si, followed by the least and alternate between next highest and next lowest until roughly one-half of dampers have been adjusted a second time. If neces say, adjust all dampers. TABLE 69. larget Method Example The results are shown under the column labeled “After Second Round.” Do « Pitot mmeroe for each branch duct to determine the final observed airflows and Ora values: ‘Ie results are shown under the column labeled “Aster Second Round.” Note that values of Quio Fange from Lis to 1.23. See the values are all about the same, the branch airflows have been adjusted tothe correct rela- tive distribution. Using minimum Onin value determined above ('min(Qraio)"), adjust the fan speed from the original rotation rate (0) to; ‘onfmin(Qratio) 598 1pui/t.23 = 486 ype The last columns in Table 4-9 show the resulting meas- ured value after moditying the tan speed. Ihe last col- umn shows the percentage of excess air in each branch. Note that the Qos 1 only 1.4%, a very small level of wasted airflow, Goal Measured ‘Computed Branch Qsost Qopent ‘SP ratio Rank k SPrgoat_ | SPhinrge A 400 328 118 082, 6 1.00 178 216 2A 300 620 1s 132 4 008 oe7 102 38 400 32 144 ost 5 099) 176 2 40 400, 584 1.88 139 3 097 087 099 ee 700 ees 188 180 2 208) O87 0.08 6D 400 1030 192 257 1 082 028) ost Fan 2400 1.40 Qiargoel 2400 ‘acim (CraroperI temgoat © MIN Orato Qanopent | 3357 ‘acim tania! 2026 ‘acim ‘TABLE 4-9 (Cont), Targot Method Example Measured after first round ‘Computed Bronoh @ Pre SPUSPimwe | SPinaw Rank k ‘SPirumyet A 503 2m 158 108 1 092 255 2A 448 1.06 1.28 0.85) 4 0.98) 1.25 38 #0 249) 124 0.04 2 096) 2.55 4c 467 148 136 094) 6 +00) 1.28 5c 496 133 154 1.06 5 099) 128 oD ay ad) 1.58 10F 3 09) az Fan 2848 1.45TABLE 49 (Cont). Target Method Example Balancing Duct Systems with Dampers 4-19 Mocourea Computed Measured [Computed Branch @ SP SPra/SPhgoa [Graton ae Orne oxses tA 400 263 150 123 aia 1.04 36% Za a0 re Ta Ta we 7 a 8 am 250 142 119 203 107 OB ac 2 128 146 121 408 102 21% 56 a5 122 tai 119 a2 00 04% 60 a a0 a0 18 200 00 00% Fan 2878 24 14% Notes: 44 IWdel’chik. LE: Handbook of Hveraulic Resistance Coefficients of Local Resistance and Friction, “US 1) The example does not include measurement error in determining the otatic presaure and volumetiie flow at ‘each measurement point 2) Uhe changing ofthe fan speed assumes the sheaves are available to provide the desired fan speed! 4.11.3 Advantages of Target Method. The Tarzet Method hus the following advuntages over the Modified Basic Method: 1. Iedocs not requir that the fan speed be adjusted before adjusting dampers. 2. It may requir only 1.5 rounds of damper adjustment ang should seldom, if ever, require more than two full rounds ‘The disadvantage of this method compared to the Modified, Basic Method is that it requires more mathematical operations, Dodrill# found remarkably good distributions using the ‘method to achieve two challenging distributions in an experi- ‘mental 7-branch aystem. Only onc and onc-half rounds of adjustments were required to obtain Quises values less than 2° for every test. The results were comparable to those found by Geiger for the “Static Pressure Ratio method.” REFERENCES 4.1 Gufley, S.E.: Airflow Redistribution in Exhaust Ventilation Systems Using Dampers and Static Pressure Ratios. Appl Occup. Environ, Hyg., £(3):168-177 (193), 42. Caplan, K.: Balance with blast gates ~a precarious ‘balance. Heating Piping/Air Conditioning, February (2003) 43° Crowder, J.W. Londermilk, KJ: Balancing of Industrial Ventilation Systems. Control Technology News, Vol 32, NO, | (1982), Atomic Enerey Commission. AEC-TR-6630 (1972) 4.5 Balasubramian, V Effectiveness of the common ‘method of adjusting exhaust ventilation system dampers. Master's Thesis, West Virginia University (2008), 4.6 Wang, Lt: Investigation of Measurement Error and, Possible Shortcut Methods in Determining Mean Velocity in Ducts, Master's Thesis, University of Washington (1997), 4.7 MeLoone, H.E.: Guflev. S.E.: Curran. J.C.: Effects of Shape, Size, and Air Velocity on Entry Loss Factors of Suction Hoods. Am. Ind. Hyg. Assoe. 1. $4(3)87-04 (1993), 48 Jorgcnsen, R. (editor): Pan Cngincesing, Dighthy Edition. Buffalo Forge Company (1983) 49 Guffey, S.E.; Spann, J.G: Experimental [nvestigation of Power Loss Coefficients and Statie Pressure Ratio ‘in an Industrial Exhaust Ventilation System, Am. Ind Hyg. Assoc. J, 60(3):367-376 (1999), 4.10 Gulley, S.E. Airflow Distribution in Exhaust ‘Ventilation Systems. Am Ind, Hyg. Assoc. J 52(3)93-106 (1991). 4.11 Dankill, M.W., Eaprasnnvnal Validation uP Un “Targut Hood Static Pressure” Balancing Method for Exhaust ‘Ventilation Systenis. Master's Thesis, West Viagista University (2004), 4.12. Geiger, H.M.: Evaluation ofa Proposed Static Pressure Ratio Balancing Method. Master's Thesis, University ‘of Washington (1999).