1、2008 ASHRAE 387ABSTRACTThe overall system efficiency (wire to water efficiency) ofa variable speed pumping system (VSPS) is determined bydrive, motor and pump efficiencies. All three component effi-ciencies vary with the pump operating conditions. A 15 hpexperimental VSPS was developed to measure th
2、e componentand system efficiencies over a wide range of operating condi-tions. The test results demonstrate the significant energysavings of a VSPS but at partial load condition the system effi-ciency is significantly lower than the efficiency at full load ornear full load conditions. Especially, wh
3、en the actual flow rateis less than 60% of the full load flow rate, the system efficiencydecreases more rapidly with decreasing flow rate. The systemefficiency also displays a strong dependency on system flowrate. Based on test results, a correlation was developedbetween the non-dimensional system e
4、fficiency and the non-dimensional flow rate. This correlation provides a simplemethod for calculating VSPS efficiency for a wide range ofpart-load operating conditions and is recommended as analternative to assuming full load efficiencies remain constant. INTRODUCTIONVariable speed drives or variabl
5、e frequency drives(VSDs or VFDs) are widely used in HVAC pumping systems.The biggest advantage of VSD system is the energy savingpotentials at part-load conditions. The affinity laws havebeen used to estimate the energy savings of variable speedpumping systems. Practically it is not possible to achi
6、eve allthe energy savings potential predicted by the affinity laws.The actual energy savings are affected by many factorsincluding departure from the idealized system curves, controlmethods, and overall system (wire-to-water) efficiency atpart-load conditions. Among these three factors, the overalls
7、ystem efficiency is often not given enough considerationbecause the complicity of the issue is not fully understood.The overall system efficiency is the product of the effi-ciencies of the three components in a variable speed pumpingsystem: the motor, the drive and the pump. Drive efficiency atfull
8、speed and full torque condition is normally provided bythe drive manufacturer. (Walski et al. 2003) However, the effi-ciency data of motor and drive efficiencies at different motorspeeds and torque conditions are not well documented. Previ-ous studies (Gao et al. 2001, Domijan et al. 1997) show that
9、 themotor and drive efficiencies vary dramatically with motorcontrol frequency (speed) and motor output torque. Needlessto say that variable speed drives can reduce energy consump-tion and operating costs significantly at part load conditions,however, in order to estimate the energy savings accurate
10、ly thesystem and component efficiencies over the full range of oper-ation are required. It is critical to understand pumping systemefficiency at lower speeds, especially for applications wherepart load conditions dominate (Kavanaugh et al. 2004, Rookset al. 2003).EXPERIMENTIAL SYSTEMAn experimental
11、variable speed pumping system wasbuilt to measure the real pump system efficiency over fullrange of operation. Figure 1 is the diagram of the experimentalpumping system. The system consisted of three major compo-nents: a motor-drive system, a pump, and a piping system. Themotor is 15 hp (11.2 kW), 1
12、750 rpm, 460 VAC-3 phase induc-tion, energy-efficient model. The drive is a pulse width modu-lating (PWM) drive with a switching (carrier) frequency thatcan be varied from 4 kHz to 16 kHz. The pump is an end-Development of a Correlation for System Efficiency of a Variable-Speed Pumping SystemXingshu
13、n Gao, PhD, PE Sally A. McInerny, PhD, PE Stephen P. Kavanaugh, PhDMember ASHRAE Member ASHRAE Fellow ASHRAEXingshun Gao is a senior engineer at TransAir Manufacturing Corporation, Dallastown, PA. Sally A. McInerny is a professor and directorof undergraduate programs in the Department of Mechanical
14、Engineering at The University of Alabama at Birmingham, Birmingham, AL.Stephen P. Kavanaugh is a professor in the Department of Mechanical Engineering, The University of Alabama, Tuscaloosa, AL.NY-08-0442008, American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc. (www.ashrae
15、.org). Published in ASHRAE Transactions, Volume 114, Part 1. For personal use only. Additional reproduction, distribution, or transmission in either print or digital form is not permitted without ASHRAEs prior written permission.388 ASHRAE Transactionssuction centrifugal pump rated for 15 hp (11.2 k
16、W) at1750 rpm nominal pump speed. A piping system is linked tothe pump that is designed for a full-load flow rate of 500 gpm(1900 Lpm) and 70 ft (21 m) of water head. The piping systemhas four branches between the main supply and return pipes.The flow rate and the pressure loss of the piping system
17、can bevaried by adjusting the control valves on branches and mainsupply and return pipes.The three-phase voltage and current into and out of thedrive were measured using voltage dividers and clamp-oncurrent sensors. The sensor outputs were recorded by a PCbased data acquisition system with a samplin
18、g rate of 10 kHz.Details on the voltage and current measurements and calcula-tion of electrical power can be found in previously publishedresearch by McInerny et al. (2001). The motor speed andtorque were measured and used to calculate the motor outputpower (pump input power). The motor shaft speed
19、wasmeasured using an optical tachometer; the shaft torque wasmeasured using an in-line rotary torque meter. System flowrate and pressure head were also measured so that pump outputpower can be determined. The flow rate was measured on themain return pipe using a turbine flow meter. Two pressuretrans
20、ducers were used to measure the pressure differentialbetween the suction and discharge of the pump.TEST RESULTSTest points are listed in Table 1. The motor controlfrequency varied from 24 Hz to 60 Hz (nominal pump speedfrom 700 to 1750 rpm) with an increment of 5 Hz (except from24 to 30 Hz). Tests w
21、ere conducted at each incremental pumpspeed from the maximum possible flow rate (open throttlecondition) at that pump speed down to about 50% of the maxi-mum possible flow rate, in increments of approximately50 gpm (190 Lpm). The selected test points covered a widerange of typical pump operations. T
22、he drive switchingfrequency was set to 4 kHz, because the motor/drive efficiencyis higher at lower switching frequencies (Gao et al. 2001,Mohan et al. 1995).Test results are given in Table 2. For each test condition,which is determined by pump speed and flow rate, the powerinput and output of the ea
23、ch components were calculated andlisted. The component efficiencies and combined system effi-ciencies were also determined and presented in the Table 2.The total system efficiency data are presented in Figure 2.The test results reveal the behavior of the motor-drive effi-ciency over a wide range of
24、operating conditions. The highestdrive efficiency is 95% and it occurred at full speed (60 Hz/1750 rpm). The lowest drive efficiency is 85% and it occurredat the lowest tested motor speed (24 Hz/700 rpm) and thelowest flow rate (lowest motor torque).At a fixed motor control frequency (pump speed), t
25、hecombined motor-drive efficiency decreases with decreasingflow rate. In other words, at same pump speed, the motor-drive efficiency decreases when system throttles. Thedecrease is more significant at lower pump speeds. The com-bined motor-drive efficiency also decreases when the motorcontrol freque
26、ncy is reduced while the flow rate is held con-stant. The motor-drive efficiency is low when the motor con-trol frequency and output torque are low (low flow rate andlow pressure conditions). At the maximum possible systemflow rate (all the valves fully open), the motor-drive efficiencyFigure 1 Expe
27、rimental variable speed pumping system.Table 1. Pump Speed and Flow Rate Test PointsMotor Control FrequencyNominal Pump Speed rpmPump Flow Rate inGallons per Minute60 Hz 1750 500 450 400 350 300 X55 Hz 1604 466 450 400 350 300 25050 Hz 1458 421 400 350 300 X 20045 Hz 1313 375 350 300 250 200 15040 H
28、z 1167 337 300 250 200 150 10035 Hz 1021 300 250 200 150 100 X30 Hz 875 250 200 150 100 X X24 Hz 700 200 150 100 X X XFigure 2 Overall system efficiency vs. flow rate.ASHRAE Transactions 389decreases with decreasing pump speed. The efficiency de-creases from 89.2% at 60 Hz and 500 gpm (1900 Lpm) to6
29、3.1% at 24 Hz, 200 gpm (760 Lpm). The behaviors of thedrive and motor efficiencies are in accordance with the previ-ous findings. (Gao et al. 2001) All the data points presented intable 2 were tested with 4kHz drive switching frequency. Forhigher switching frequency, the trend is expected to be thes
30、ame but the efficiencies are lower because of the highermotor-drive losses.The pump efficiency does not show the same trend as thedrive and motor. It follows the pump curves. When running atconstant speed, pump efficiency does not change significantlyuntil the flow rate is reduced substantially. Amo
31、ng the testingpoints which have the same flow rate, the one has lower pumpTable 2. Selected Pumping System Efficiency Test Results (4 kHz Switching Frequency)Motor Control FrequencyPump SpeedrpmFlow RatePMPressureft H2OTorquein lbsVSD PinwattEfficiencyVSDEfficiencyVSD-MotorEfficiencyPumpEfficiencyWi
32、re to Water24 Hz 714.9 200.9 10.2 58.3 781.6 87.1% 63.1% 78.5% 49.5%715.4 151.4 11.5 50.3 702.9 86.3% 60.5% 77.4% 46.8%715.9 101.0 12.3 41.5 622.1 85.2% 56.6% 66.7% 37.7%30 Hz 892.6 251.3 15.7 91.9 1338.0 89.5% 72.5% 76.5% 55.5%893.2 200.9 17.5 82.2 1232.4 89.0% 70.5% 76.5% 53.9%893.7 149.8 18.9 71.
33、7 1102.9 88.8% 68.7% 70.3% 48.3%894.5 101.2 19.5 60.7 972.1 87.8% 66.0% 58.0% 38.3%35 Hz 1040.5 295.5 21.1 128.4 1953.5 90.0% 80.9% 74.5% 60.3%1041.1 250.1 23.2 120.0 1837.8 89.6% 80.4% 73.9% 59.4%1041.8 200.4 24.9 107.3 1682.0 89.3% 78.6% 71.0% 55.8%1042.6 150.4 26.0 93.1 1494.9 88.7% 76.8% 64.2% 4
34、9.4%1043.5 102.2 26.3 79.3 1314.1 88.5% 74.5% 51.7% 38.5%40 Hz 1188.4 337.3 27.6 169.4 2824.0 92.3% 84.3% 73.7% 62.2%1188.7 301.1 29.6 158.4 2664.8 92.3% 83.6% 75.4% 63.0%1189.4 251.9 31.7 144.5 2461.3 92.1% 82.6% 73.9% 61.1%1190.3 201.5 33.3 129.4 2226.1 92.0% 81.8% 69.4% 56.8%1191.2 149.9 34.2 113
35、.1 1982.8 91.7% 80.4% 60.5% 48.6%45 Hz 1335.5 376.0 34.8 212.4 3918.6 90.4% 85.7% 73.5% 62.9%1335.9 349.9 36.4 204.2 3780.6 90.5% 85.4% 74.3% 63.4%1336.7 299.6 39.0 187.7 3495.9 90.3% 84.9% 74.2% 63.0%1337.6 252.7 41.1 172.1 3224.2 90.2% 84.5% 71.8% 60.6%1338.5 200.3 42.6 153.6 2913.1 90.2% 83.5% 66
36、.1% 55.2%1339.7 150.0 43.3 134.2 2597.2 89.6% 81.9% 57.5% 47.1%50 Hz 1482.7 421.9 42.4 262.0 5242.7 93.6% 87.7% 73.3% 64.3%1482.9 401.0 43.9 254.2 5096.6 93.6% 87.5% 74.3% 65.0%1483.9 349.4 47.1 236.3 4763.9 93.6% 87.1% 74.7% 65.0%1484.6 299.4 49.7 219.7 4425.3 93.4% 87.2% 72.6% 63.3%1486.7 201.4 52
37、.6 181.7 3736.0 93.2% 85.5% 62.4% 53.4%55 Hz 1629.4 466.0 51.0 312.4 6797.8 92.4% 88.6% 74.4% 65.9%1629.5 450.6 52.0 308.8 6721.7 92.4% 88.6% 74.2% 65.7%1630.1 400.7 55.6 291.5 6371.2 92.4% 88.2% 74.7% 65.9%1631.4 350.4 58.7 271.4 5952.6 92.3% 88.0% 74.0% 65.1%1632.4 302.2 63.7 251.5 5517.6 92.4% 88
38、.0% 74.7% 65.8%1633.4 250.0 62.8 231.3 5105.5 92.3% 87.5% 66.2% 58.0%60 Hz 1776.3 506.8 60.6 370.6 8735.4 93.5% 89.2% 74.3% 66.3%1776.5 450.6 64.8 351.2 8306.0 93.7% 88.9% 74.6% 66.3%1777.8 401.4 68.6 331.1 7848.0 93.9% 88.7% 74.5% 66.1%1778.4 349.4 71.8 307.0 7312.6 93.0% 88.3% 73.2% 64.7%1780.0 30
39、2.1 73.3 287.2 6850.8 95.0% 88.3% 69.0% 60.91390 ASHRAE Transactionsspeed has higher pump efficiency. This is exactly the oppositeof the motor-drive efficiency. For example, with 200 gpm flowrate, at 50 Hz motor control frequency (1458 rpm), the motor-drive efficiency is 85.5% and the pump efficienc
40、y is 62.4%; at30 Hz motor control frequency (875 rpm), the motor drive effi-ciency is 70.5% but the pump efficiency is 76.5%. Therefore,the overall system efficiency, as the product of the motor-driveefficiency and pump efficiency, varies less drastically with themotor control frequency (pump speed)
41、 at the same flow ratecomparing to the component efficiencies. The overall systemefficiencies at the two mentioned conditions are 53.4% and53.9% respectively.The over all system efficiency (wire-to-water efficiency)vs. flow rate curves are shown in Figure 2. All the test pointsthat have the same mot
42、or control frequency (pump speed) areplotted on one same curve. The overall system efficiencydisplays a strong dependence on the system flow rate: theoverall system efficiency decreases with decreasing flow rate.Furthermore, the data points distribute within close vicinity ofa trend curve running in
43、 middle of the data points with a fewpoints lying outside a rather narrow envelope. This indicates apossibility that the wire-to-water efficiency can be presentedas a function of the system flow rate and independent of thepump speed. Excluding the outlaid data points, an exponentialregression curve
44、can be generated to represent the relationshipbetween the system flow rate and the overall system efficiency.See Figure 3. In order to make this equation generic, the vari-ables are presented in non-dimensional forms. The non-dimensional flow rate QEis the ratio of the actual flow rate andthe full l
45、oad flow rate. The non-dimensional overall systemefficiency Eis the ratio of actual efficiency and the full loadefficiency. It can be seen from Figure 3, the overall system effi-ciency is close to the full load efficiency when the flow rate isabove 60% of the full load flow rate. However, when thesy
46、stem flow rate decreases further, the overall system effi-ciency decreases significantly.(1)(2)(3)The data points that lay below the regression curve occurat the maximum possible flow rates (full open conditions) atrespective pump speeds. (Especially at lower speeds.) Forexample, at 200 gpm, overall
47、 system efficiency is 49.5% for24 Hz motor control frequency (700 rpm), while the overallsystem efficiency at higher motor control frequencies isaround 55%. Here, the motor-drive efficiency starts to domi-nate the overall system efficiency and a modest increase insystem efficiency can be obtained wi
48、th an increase in pumpspeed and adding some throttle. In reality, it is not likely tohave a full open condition at lower pump speeds (partial loadconditions), due to the increasing of the throttling (i.e., morevalves are closed at lower pump flow rate conditions). To meetthe required pressure head a
49、t the flow rate, the pump speedmust be increased.For pumping systems in which these extreme conditionsare not likely to occur, the regression curve can well cover therealistic operating conditions. However, it is important tounderstand that the “head area” (also known as control area)of a particular variable speed pumping, may exceed testedpump operating range due to different piping designs and loadconditions. (Hegberg 2003) Further studies are required toreveal the behavior of component efficiencies over all thepossible operating conditions.SUMMARYThe wire-to-water efficiency of a variab
