1、4673 A Switched Reluctance Motor in a Variable-Speed Pumping Application Horacio G. Vasquez, Ph.D. Student Member ASHRAE Joey K. Parker, Ph.D. Timothy A. Haskew ABSTRACT : A variable-speed pumping system based on a switched reluctance motor (SM) and a centrifugal pump was devel- oped, and it is inte
2、nded to be used in heat pump or similar applications. A ground-source heat pump (GSHP) uses a centrifugal pump to circulate water in an underground loop using water as a means to transfer heatfi-om theground to room air inside a building or vice versa. SRM. are appearing as alternative actuators in
3、many engineering applications because of their simple construction, reliability, and low manufacturing and maintenance costs. In general, an SM, its power convertel; and control strategy have to be adequately designed or selected and integrated in order for the entire system to operate efficiently a
4、nd satisfactorily. The special application ofan SMin thisproject was to drive a centrifugal pump required to operate efficiently at all speeds-in partic- ulal; at low and medium speeds. The SRM-based variable- speedpumping system was experimentally tested and demon- strated potential to save energy
5、in central system GSHP or similar applications. INTRODUCTION A central system GSHP consists of a pump, underground heat exchanger, supply and return pipes, and several air- handling units (AHUs), as shown in Figure 1. The under- ground piping system is used as a means to transfer heat from the room
6、air to the ground or vice-versa (Rafferty 1997). In some installations, additional loops and heat exchangers are installed to heat utility water. AHUs are turned on or off inde- pendently of each other, and a variable-speed drive (VSD) is required to maintain a setpoint differential pressure across
7、the AHUs and also to save energy and avoid unnecessarily high pressure and losses in the piping system. Otherwise, valves, which always waste energy, are used to regulate the water flow rate. In addition, to justify the use of a variable-speed drive for the pump, the friction losses in the rest of t
8、he piping system must be similar to or greater than the setpoint differential pres- sure across the AHUs when the pump is running at full speed. The justification of a variable-speed drive for the pump must also be based on the energy savings that can be achieved due to operation of the pump at spee
9、ds lower than full speed when the pumping system demand changes. This situation commonly occurs in central system GSHP systems; hence, a variable-speed pump is almost always required in such appli- cations. In general, a VSD motor pump system is designed to liRetum Figure 1 Central system ground-sou
10、rce heat pump. Horacio G. Vasquez is a lecturer at the University of Texas-Pan American, Edinburg, Tex. Joey K. Parker is an associate professor in the Department of Mechanical Engineering and Timothy A. Haskew is an associate professor in the Department of Electrical and Computer Engi- neering, Uni
11、versity of Alabama, Tuscaloosa, Ala. 02004 ASHRAE. 67 satis maximum heating and cooling loads in a GSHP appli- cation, but most of the time the system operates at low or medium loads due to reduced building occupancy andor favorable atmospheric conditions. It is ideal that the variable- speed drive,
12、 motor, and pump operate with high efficiency at low and medium speeds to maximize energy savings. The most common variable-speed drives (VSDs) used in heat pump and similar applications are the type called variable- frequency drives (VFD), used with conventional alternating- current induction motor
13、s (AC IMs). VFDs and AC IMs are characterized by having high efficiency when operating near their rated speed and load; however, their combined efficiency starts decreasing substantially at speeds below 50% of the motor rated speed (Bernier and Bourret 1999; Casada et al. 2000). When oversized IMs a
14、re selected, this efficiency draw- back becomes more significant (Henderson et al. 2000). Kavanaugh and McInerny (2001) determined that drive- motor-pump efficiency data are not widely available at low speeds and that additional research must be performed to establish variable-speed pump demand at l
15、ow loads and speeds. Kavanaugh and McInerny (2001) also concluded that more than 50% of annual energy consumption by the central system GSHP in a school facility occurs when the building is unoccupied and when the pump is operating at low speeds. Therefore, a more efficient, economical, and practica
16、l VSD- motor system alternative will provide important contributions toward the development of more efficient heat pumps and pumping systems that operate similarly. The main goal of this research was to determine the opportunities and benefits that switched reluctance motors (SRMs) could contribute
17、to solve this problem. Therefore, an SRM was modeled, simulated, implemented, controlled for variable speed, and experimen- tally tested, driving a centrifugal pump in a closed-loop piping system like that in a central system GSHP application. BACKGROUND Several studies have demonstrated that variab
18、le-speed drives save considerable energy when used in central system GSHPs. A few studies have demonstrated the use of SRMs to drive pumps at constant speed and at particular loads, as in a hydraulic system or in a fuel delivery system. Nevertheless, it was determined that there are not reported stu
19、dies of SRMs used in central system GSHP applications, where vanable- speed drives operating at high efficiency at low and medium speed for a long number of hours are required for the centrif- ugal pumps. Kavanaugh and McInerny (2001) addressed a study of pumping options for the air-conditioning sys
20、tem of a 72,000 fi2 (6689 m2) school facility using four ground-source heat pump (GSHP) arrangements. The heat pump options consisted of the following arrangements: (a) a decentralized system with multiple individual heat pumps throughout the building, (b) a central system with a variable-speed pump
21、, (c) a system with a constant-speed pump, and (d) a system with primary and secondary pumps. It was indicated that a well- designed ground loop does not require a high water flow rate to effectively transfer or absorb heat from the ground, which also implies operation with lower head losses. To det
22、ermine the number of hours that the pumps were required to operate for each of the heat pump arrangements, the piping loop was designed, the pumps were specified, and the pump demand based on building load was computed. Bin weather data were used with pump power demand to compute the annual energy c
23、onsumption. The decentralized system used circulator pumps running at constant speed and controlled using an odoff method; therefore, the energy consumption was computed based on the circulator rated power and the number of hours at work. For the central system, the energy consumption was more diffi
24、cult to determine, and a correlation between pump power consumption and water flow rate, power consumption at full load, and water flow rate at full load was determined. Such correlation was convenient because the relationship between building load and water flow rate was established. Consequently,
25、the total VSD-pump system energy consump- tion was computed. The authors concluded that the arrange- ment with multiple decentralized units was the most energy- efficient system of all four options, consuming approximately 44.7 x lo6 Btu (13.1 x lo3 kwh) annually. The central system GSHP with variab
26、le-speed pump followed the decentralized system in efficiency, using 44% more pumping energy than the decentralized system. The other two heat pump arrangements were much less efficient. The decentralized system is more efficient because of the long periods when the individual units are turned off d
27、uring unoccupied building hours and because of the low-pressure head required to operate them. However, the decentralized system requires more bores than the central system for the groundwater loops. The authors indicated that, in general, for a central system GSHP, the VSD-pump unit must operate at
28、 low load and speed during large numbers of hours in buildings that are occupied for less than 50 hours per week. It was mentioned that other studies have reported that VSD-motor system efficiency using a conventional induction motor is much lower when torque is less than 25% of full load (Gao et al
29、. 2001). Jones and Barrer (1999) presented a study of energy savings by using variable-frequency drives (VFDs) in open- loop systems to substitute odoff systems and throttled control systems. It was explained that an open-loop system is like a pump filling a tank opened to the atmosphere. It was det
30、er- mined that in open-loop systems, when significant and domi- nant static load is present, VFDs could use more energy than constant-speed drives operating odoff to fill or empty a tank, for example. The constant-speed pump in an open-loop system often operates at a high-efficiency condition. There
31、- fore, VFDs were not recommended in open-loop systems to substitute a pump system that operates with an odoff control scheme and at constant speed. This occurs because the fixed static head is dominant, and the savings due to VFDs are concentrated in the dynamic load and, in such case, they are off
32、set by the lower efficiency of the pump, motor, and drive at 68 ASH RAE Transactions: Research lower speeds. On the other hand, it was explained that a differ- ent situation occurs with VFDs used to substitute throttled pump systems. Throttled control always wastes energy, which could otherwise be s
33、aved with VSD-pump systems. Several charts were presented to estimate the payback period of using VFD to substitute a throttled control in a pumping system. It was shown that the higher the operating static head and the higher the flow rate, the longer the payback period of VFD systems. In other wor
34、ds, energy savings are more significant in systems that operate at less than 60% of design flow rate, at low static head, and for a large number of hours. For such systems, the payback periods are one or two years. These results agreed with the considerable savings achieved with VFDs used in closed-
35、loop systems. Carlson (2000) presented several considerations to justi the use of VSDs in pumping systems. The author used several plots and affinity laws to explain such considerations with simple examples. It was indicated that the most common errors in computing energy savings with VSDs occur by
36、not considering the correct efficiency changes in the pump, motor, and drive at different flow rates, speeds, and loads. A method to accurately estimate energy savings by using VSDs in pump- ing systems was addressed in a paper. Yamai et al. (2000) designed a 3 hp (2.2 kW), 12/8 SRM used in a hydrau
37、lic pump unit, where fast response and high efficiency were required at low-speed high-torque operating conditions. A three-phase 12/8 SRM was preferred over a three-phase 6/4 SRM because it produced less acoustic noise. The SRM operated 90% of the time at high torque and about 300 rpm while maintai
38、ning the required pressure in the system. Energy savings of more than 50% were obtained after the SRM was implemented as a variable-speed drive with a fixed displacement pump, substituting for an induction motor running at constant speed and driving a variable displacement Pump. Ferreira et al. (1 9
39、95) designed and implemented a 5 hp (3.7 kW), 270 V, 10000 rpm, 8/6 SRM as a variable-speed drive for a pump in a module of a gas turbine engine fuel deliv- ery system. The torque load for the motor was constant at about 9.7 in:lb (1.1 N.m) up to 5000 rpm and then linearly increased to 35.4 in:lb (4
40、 N.m) at 10000 rpm, and the motor was required to accelerate from O to 10000 rpm in two seconds. At full speed and load, the SRM inverter efficiency was about 84% and the SRM efficiency was 94%, using either an electronic position sensing (EPS) or a resolver, which means that the inverter-SRM system
41、 had an efficiency of about 82%. Metwally and Anis (1 996) used a three-phase, 4 hp (3 kW), 12/10 SRM to drive a water pump at a fixed operating speed of 1350 rpm. A photovoltaic array was used with a battery and voltage regulator to provide the electric energy for the SRM. It was mentioned that the
42、 DC-link voltage of the power supply changed with solar irradiation, but a minimum of 84 V was ensured most of the time. An SRM was selected because other types of motors were considered to have more complex inverters, and they were more expensive or not appro- priate for the variable DC-link voltag
43、e of the power source available. DESIGN OF A CLOSED-LOOP PIPING SYSTEM TO TEST THE SRM PUMP UNIT As a first step, in order to study the operation of the SRM- pump unit, a closed-loop piping system was designed and constructed. An SRM and a conventional AC induction motor (IM) of the same size and wi
44、th similar ratings were used to drive the same pump with the objective of studying and comparing the motor performances at different speeds. A 6/4 SRM rated at 2 hp (1.5 kW) and 1500 rpm was the component chosen first. A constraint to design the PVC piping system was to provide realistic operating c
45、onditions for the pump with a total head loss in the 30-80 ft (9.1-24.4 m) of water range when operating at rated speed. The piping system was designed to have approximately 50 fi (15.2 m) of water head loss and 100 gpm (6.31 L/s) of flow when the pump operated at 1500 rpm. Since the pump curves are
46、, in general, given by the manufacturer at 1750 rpm, the affinity laws were used to select a pump rated at 1750 rpm, 115 gpm (7.26 L/s) of water flow rate, and 70 ft (2 1.3 m) of water head loss, assuming lower pump efficiency at 1500 than at 1750 rpm. A centrifugal pump with an impeller diameter of
47、 8.2 in. (32.3 x 10-3m) rated at 1 13 gpm (6.2 L/s), 67 ft (20.4 m) water head loss, with efficiency of 66% at those operating conditions, was chosen. Under these rated conditions, the pump required a 3 hp (2.2 kW) motor. By additional computations, it was determined that the character- isticsforthi
48、spumpwere96.9gpm(5.32L/s)and49.3 ft(l5m) of water head loss at 1500 rpm. Using 59% pump efficiency, a two-horsepower motor was required at 1500 rpm. Conse- quently, this pump and the chosen two-horsepower SRM match power ratings at 1500 rpm. At the same time, realistic head loss and flow rate condit
49、ions, as in a central system GSHP, were obtained at the motor rated speed of 1500 rpm. A three-phase AC IM rated at 2 hp (1.5 kW), 50 Hz, and 1440 rpm with frame 90L, the same as the frame of the 6/4 SRM, was acquired. A commercial-type programmable AC variable-frequency drive (VFD) was also acquired for this IM. The VFD was rated at 3 hp (2.2 kW) and 60 Hz, so that the required 2 hp (1.5 kW) power was available at 50 Hz. The VSD-IM and inverter-SRM units were compared driving the same centrifugal pump. Figure 2 presents the piping system with the pump, motor and inverter, tank, venturi flow
