1、OR-05-7-3 A Methodology to Evaluate Pumping Energy Consumption in GCHP Systems Alain Sfeir Student Member ASHRAE Michel A. Bernier, PhD Member ASHRAE Thibault Million Alexandra Joly ABSTRACT operate in several modes, ranging from constant flow rate to A methodology to evaluate the energy consumption
2、 of centralized and decentralized pumping systems in ground- coupled heatpump (GCHP) systems ispresented. The analysis covers thejrst costs of the piping (interior and exterior) and of the circulating pump, as well as annual pumping energy costs. The methodology is applied to two buildings represent
3、ing two distinct buildingiclimate combinations. Results show that when the cost ofthe pumps and their installation are added to the life-cycle energy cost of 20 years ofpumping, centralized systems for both buildings have total pumping costs that are lower than decentralized systems. It is also show
4、n that the annual energy-weighted average wire-to-water pump e ciency of centralized systems is much better than for decen- tralized systems. Despite this pumping cost advantage, the overall costs (interior piping + exterior piping + pumping- related costs) ofcentralizedsystems are higher than for d
5、ecen- tralized systems. This is primarily due to the high interior piping costs of centralized systems. INTRODUCTION A typical GCHP system consists of heat pump(s) to heat/ cool the building, a ground heat exchanger (GHX) to collect/ reject heat to the ground, and pump(s) to circulate a thermal flui
6、d between the heat pumps and the ground heat exchanger. As shown schematically in Figure la, there are two schools of thought when it comes to fluid pumping in GCHP systems. The first one is referred to as “centralized” pumping, whereby a single pump circulates the thermal fluid from the borefield t
7、o the heat pumps. The circulating pump can be chosen to variable flow rate, using variable frequency drives (VFDs). The second type of system is “decentralized” (Figure lb). In these systems, each heat pump has an individual pump-GHX loop. Each type of system has its advantages and drawbacks. Centra
8、lized pumping systems have better nominal wire-to- water pumping efficiencies than decentralized systems, as these latter systems often use low-efficiency circulators. In addition, the GHX of centralized systems tend to be shorter than their decentralized counterparts. This is because central- ized
9、systems experience smaller peak loads due to load diver- sity in the building. This piping cost advantage for the borefield is often counterbalanced by simpler piping arrange- ments for decentralized systems, which lower interior piping costs and reduces the theoretical pumping power (Outside the Lo
10、op 1999, 2000). Pumps in decentralized systems usually operate in onoff mode based on whether the heat pump is operating or not. Thus, during off-peak conditions, decentral- ized systems will tend to be less energy-intensive than central- ized pumping systems where the central pump(s) is still runni
11、ng, albeit at a reduced speed in most cases. As can be seen from this list ofpros and cons, it is not clear which system should be selected for a given application. This paper addresses this issue by presenting a methodology that could be used to analyze each system and compare their life- cycle cos
12、ts (LCC). The analysis covers the first costs of the piping and of the circulating pump, as well as annual pumping energy costs. An application of this methodology is also presented. _ _ - - Alain Sfeir is a graduate student and Michel Bernier is a professor in the Dpartement de gnie mcanique, cole
13、Polytechnique de Montral, Montral, Qubec, Canada. Thibault Million and Alexandra Joly are students in the Dpartement de gnie civil et urbanisme, INSA-Lyon, Lyon, France. 774 02005 ASHRAE. Central Zone Figure 1 Schematic representation of centralized and decentralized pumping systems. ASHRAE Transact
14、ions: Symposia 71 5 REVIEW OF PREVIOUS STUDIES In a recent ASHRAE-sponsored survey (Cane et al. 1998), it was found that installed pumping power represented anywhere from 0.04 to 0.21 HP per ton (0.0085 to 0.045 kWelectkWtherm) of installed heat pump capacity. If one assumes that the sum of the annu
15、al equivalent full-load heating and cooling hours is 1500 hours and that the heat pumps have an average COP of 3.5, then pumping energy consumption for the surveyed systems may range from 15% to 48% of the total (heat pumps + circulating pumps) energy consumption when the circulating pumps operate c
16、ontinu- ously. In a similar analysis, Kavanaugh (1 996) has shown that a circulating pump can represent 45% of total energy consumption. Thus, contrary to popular belief, circulating pumps may represent a significant portion of the total energy consumption of GCHP systems. In their book, Kavanaugh a
17、nd Rafferty (1 997) introduced a set of guidelines to limit pumping power in GCHP systems. It is suggested that pump motor size should fall in the range of 10 to 13.3 kW per 350 kW of refrigeration (7.5 to 10 HP per 100 tons). However, no indications are given on the corre- sponding yearly energy co
18、nsumption. Results of an ASHRAE research project (Kavanaugh et al. 2003) pertinent to the present article are also noteworthy. Using simple bin calculations, this study has shown that small on-off circulator pumps in decentralized systems are the lowest pump energy options for low to moderate occupa
19、ncy buildings ( 60 hours a week) and were a close second to on-off circulator pumps for build- ings with low to moderate occupancy. Before presenting the methodology to evaluate piping and pumping costs, it is important to review the available data concerning the various efficiencies of a pump-motor
20、-VFD arrangement. PUMPING POWER In its most general form, a pump is composed of three major components: (1) the pump itself, (2) the electric motor driving the pump, and (3) a variable-frequency drive. Each component has an intrinsic efficiency, and the various power requirements for each component
21、are given by the following equations (Bernier and Bourret 1999; Bernier and Lemire 1999). The theoretical pumping power (often referred to as the water horsepower) is defined by Ptheo = Q SG in si units and 102 wG in I-P units, 3960 ?the0 = denominators are simply conversion factors to obtain units
22、of kilowatts and horsepower, respectively. The amount of power required at the drive shaft of the pump, Pshaft, is given by ?the0 ?shaft = - % where qp is the pump efficiency. The electric power required to feed the electric motor, P, is given by p, = qrn (3) where qm is the efficiency of the electr
23、ic motor. Finally, the required electrical power at the inlet of a pump-motor-VFD configuration is given by (4) p. = ?the0 in qVFD qrn qp (for pumps with Ptheo 0.375 kW 0.5HP), where q vFD is the efficiency of the VFD. The denominator of Equation 4 repre- sents the so-called ?wire-to-water efficienc
24、y.? Circulators in Decentralized Systems Small capacity pumps (or circulators), which are defined here as pumps having a water horsepower less than 0.375 kW (0.5 HP), rarely use VFDs. The pump and motor efficiencies are usually combined into a single value. This value will be referred to as the circ
25、ulator wire-to-water efficiency, qc. Thus, in the case of circulators, Equation 4 becomes ?the0 p. = - in rlC (5) (for pumps with Ptheo F rc s S a, o .- 5 I O0 95 90 85 80 O 20 40 60 80 1 O0 Percentage of nominal speed, S Figure 2 Circulator wire-to-water eficiency us a function of the theoretical p
26、umping power (Water HP). Figure 3 Variable-jrequency drive eficiencies (q vFO, as a function ofpercentage of nominal speed (S). Pumps-Electric Motors-VFD in Centralized Systems In decentralized systems, the circulator pump usually stops when heat pumps are not operating. In a centralized system, how
27、ever, on-off valves are usually installed to stop water flow through idle heat pumps. In doing so, the required flow rate is reduced. In such situations, centralized pumps are often equipped with variable-frequency drives, which vary flow rate by varying pump speed (Bernier and Lemire 1999). As indi
28、cated in Equation 4, three efficiencies are therefore required to obtain the pumping power in these systems: q vFD, rim, rip. Variable-Frequency Drive Efficiencies, q vFD. As mentioned in a recent publication (Gao et al. 2000) VFDs have inherent inefficiencies. The 2000 edition of the ASHRAE Handboo
29、k presents q vFD values as a function of motor speed. These data have been reproduced in Figure 3 along with the experimental data of Gao (Gao et al. 2001). As shown on this figure, new VFDs have much better efficiencies that older ones. It is also worth noting that the experimental data (Gao et al.
30、 2001) is in good agreement with the curve suggested by ASHRAE for both the 4 and 16 kHz switching frequencies. In this work, the value of q vFD is obtained using the following equation, which is based on the curve suggested by ASHRAE for new VFDs: qvFD = 87.84 + 0.225 x S- 0.001228 x s (7) factor,
31、Fc, which accounts for decreasing efficiency of electric motors at part-load operation. qrn = qn Fc (8) The US Department of Energy presents NEMA threshold full-load nominal efficiency values for energy-efficient motors (U.S. Department of Energy 1996). These data cover the range from 1 to 500 HP an
32、d are presented for four different motor speeds (900, 1200, 1800,3600 rpm) and for open and closed motors. Only a subset of this database is used in the present study. It is presented in Figure 4. As shown in Figure 4, only motors having nominal power values in the 1 to 50 HP (0.75 to 37.5 kW) range
33、 are consid- ered. Each data point represents the average of six values (1200-1 800-3600 RPM for both the open and closed motors) obtained from the DOE data. Values of qn vary from 8 1 % for a 0.75 kW (1 HP) motor to approximately 93% for a 37.5 kW (50 HP) motor. A curve fit applied to the data of F
34、igure 4 gives 14.3 x P, 3.18 + P, q, = 79.35 + - (9) for 0.75 2 I and CAP. - I - MHP, where “Pi is the number of heat pumps in zone i. Based on these values, heat pumps with a certain nominal capacity, will be selected. For the sake of simpli9ing calcu- lations, it is assumed in this study that all
35、heat pumps in a ther- mal zone are identical. Furthermore, it is assumed that all heat pum s are controlled by a thermostat. Therefore, values of CAP, and CAP: are the sensible heating and cooling capac- ities and HDL, and CDL, are sensible loads. However, values of Diversified Design Block Load. In
36、 order to account for load diversity in centralized pumping systems, diversified design block cooling and heating loads, DBLc and DBLh, need to be calculated. The hourly values of DBLc and DBLh are determined by adding up the zone loads during a given hour j of the heating/cooling design day: 2 repr
37、esent total (sensible + latent) capacities. n n DBL; = CDL, DBLJ = c HDL, (14) i= 1 i= 1 where n is the number of thermal zones in the building. The DBLs are then calculated by taking the maximum of DBLj over the 24 hours of the design day: DBLC = max(DBL,c , forj = 1 to 24); DBL = max(DBLj , forj =
38、 1 to 24) h (15) h and finally the diversified design block load, DBL, will be the largest of these last two values: DBL = max(DBLC, DBL) (16) Hourly Heating and Cooling Loads. Hourly heating and cooling loads of each thermal zone are essential for an accurate determination of the annual pumping ene
39、rgy consumption. They are also used to size the ground heat ASHRAE Transactions: Symposia 71 9 exchanger. In this work, hourly values of the heating and cool- ing loads will be referred to as HLiJ and CL, where the i and j indexes refer to a thermal zone and to a particular hour, respectively. There
40、 are a number of commercially available software programs that can be used for this task. In this work, HLjj and CLij are evaluated using the building model (TYPE 56) of TRNSYS (Klein et al. 2000). Hourly Flow Rate zone during a given hour, denoted here by OTp is given by OT = if the heat pump opera
41、tes in cooling mode The operating time (in percent) of each heat pump in each CL, . “ CAP:NHP HL, . CAP: x NHP, OT = if the heat pump operates in heating mode J (17) Decentralized Systems. The nominal pumping flow rate of each circulating pump in zone i of a decentralized system, Qnom, , , will be d
42、etermined according to the recommended nominal flow rate per unit load, Qnom,per-un;t-ioad and the nominal capacity of each heat pump: . dec (18) . dec Qnom,i = Qnom,per-unit-Ioad X CAPnom,i Since decentralized systems operate in on/off mode, the total pumping flow rate of each circulating pump in z
43、one i at a given hour j, Qtotal,ij is given by . dec . dec Qtotai,ij = 0 when the heat pumps are not operating and .dec - . dec Qiotai,ij - Qnom,i when the heat pumps are operating. in centralized systems, Qtotalj, are obtained by Centralized Systems. The hourly values of the flow rate . cen Then, t
44、he total nominal flow rate, Q;tya, nom, is obtained by evaluating the maximum flow rate occurring during the year. (21) . are replaced by CAPnom,i in Equation 17. Two assumptions are used to determine Q;i:ali in Equa- tion 20. First, the values of OT. are based on constant heating i and cooling capa
45、cities (CAP, and CAP: ). These capacities depend, as stated earlier in connection with Equation 13, on the minimum/maximum return temperatures from the ground loop. Since heat pumps experience small capacity variations over relatively large entering water temperature variations (Kavanaugh and Raffer
46、ty 1997), it is reasonable to assume that CAP: and CAP; are constant throughout the year. One non- negligible benefit of this assumption is that hourly ground loop temperatures do not need to be evaluated. The second assumption behind Equation 20 has to do with how heat pumps operate during a given
47、hour. It is assumed here that each heat pump requires an “average“ flow rate during a given hour. This situation is perhaps best explained by looking at a simple example. Lets assume that the heating loads at a certain hour j of a two-zone building are 10 and 15 kW, respectively. Each zone has one h
48、eat pump with corresponding heating capacities of 20 and 30 kW, respectively, with nominal heat pump capac- ities of 25 and 35 kW. Then, according to Equation 17, OTij = 10120 = 0.5, OT2j = 15/30 = 0.5 and, assuming a recom- mended nominal flow rate per unit load of 0.054 Lls-kW (3 gpdton), the tota
49、l pumping flow rate for this first scenario is simply (Equation 20): . cen Qtotai = 0.5 x 0.054 x 25 x 1 + 0.5 x 0.054 x 35 x 1 = 1.62 L/s (25.7 gpmj on average for hourj. One can think of a multitude of other scenarios that could occur during hourj. For instance, it could be argued that each heat pump will be fed with its nominal recommended flow rate during the period in which it operates. If that is the case, and assuming that both heat pumps fulfill their heating duty during the first 30 minutes of hourj, then pumping flow rates will be The
