1、471 O - Climatic Impacts on Heating Seasonal Performance Factor (HSPF) and Seasonal Energy Efficiency Ratio (SEER) for Air-Source Heat Pumps Danny S. Parker Associate Member ASHRAE Matthew Lombardi Philip Fairey Member ASHRAE Bruce Wilcox Member ASHRAE ABSTRACT Within ratingprocedures established by
2、 the US. Depart- ment of EnerD, benchmarks have been established for the comparative performance of heat pumps and air conditioners. The heating seasonal performance factor (HSPF) and seasonal energy eficiency ratio (SEER) index heating and cooling season performance, respectively. Although theproce
3、- dures result in a highly desirable standard metric, the climate- related limitations of the published values must be under- stood-particularly when attempting to extend performance prediction across regions. This paper describes evaluation of climate-related variation of heat pump and air-conditio
4、ner performance. Operating seasonal eficiencies are statistically related to location-specijic winter andsummer design temper- atures and manilfacturers equipment ratings. Implications are discussed. HEAT PUMPS Residential air-source heat pumps are an increasingly popular heating system in the South
5、ern United States. Over 10 million heat pumps (HPs) are currently in use (EIA 200 I). The practical efficiency that air-source heat pumps achieve is a coefficient of performance (COP) of 2.0-3.0. To rate heat pumps in a standard fashion, a heating seasonal performance factor (HSPF) is determined tha
6、t takes into account operation under varying outdoor temperatures as well as part-load impacts (effects of running short cycles under mild condi- tions). HSPF is rendered as BtdWh so that typical HSPFs are nominally on the order of 6.8-10 BtdWh (the dimensionless value of the minimum HSPF of 6.8 is
7、COP = 1.99). HSPF is defined according to test procedures as promulgated by AM in its Standard 210/240, as well as ASHRAE Standard 116 and the DOE Test Procedure in 10 CFR, Part 430, Appendix M (AM 2003). The rateanameplate HSPF from AR1 2 1 O240 is based on the temperature in climate region IV (200
8、0-2500 heating load hours) and the minimum design heating requirement (DHR) that is a function of machine heating capacity. This selection is favorable to limit the contribution of resistance heating because it typically results in a balance point in the 17F to 25F range. Although published HSPFs ar
9、e linked to this climate, and specifically to 2080 heating load hours, it was never envisioned that this single value could be used to gener- ically predict performance for all climate locations. Given the severity of winter in much of the continental United States and the sensitivity of heat pumps
10、to the outdoor temperature, site- specific performance must vary significantly with climate. Although temperature bin data and procedures within AR1 2101240 are available to compute performance in other climatic regions (Sections A.6.2.4 and A.6.2.5 of that stan- dard), the published data available
11、for all heat pump and air- conditioning units is that of climate zone IV. Thus, although a method is available to compute HSPF and SEER in other regions, this is not done, and the information is unavailable to consumers and others without access to data on machine performance at the specific test po
12、ints. The AR1 directory of certified unitary air-source heat pumps does provide rudimentary information on heating performance for all six heating regions. The directory includes a table entitled “Heating Cost Factor,” which provides infor- mation that can be used to adjust the annual energy cost ba
13、sed on the FTC-labeled amount for Region IV for each unit that is listed to annual energy costs for the other five regions. Unfor- tunately, as will be shown within this paper, the climate clas- sifications within the ARI directorys map leave much to be Philip Fairey is director, Danny S. Parker is
14、principal research scientist, and Matthew Lombardi is a research assistant at the Florida Solar Energy Center, Cocoa, Fla. Bruce Wilcox is principal at Berkeley Solar Group, Bcrkeley, Calif. 1 78 02004 ASHRAE. desired relative to accuracy in capturing winter severity. Also, the methods used to estim
15、ate the heating system performance within the procedure itself tends to be optimistic relative to typical operating conditions, setpoints, etc. Moreover, the six climatic zones available within the ARI 21 O240 method are necessarily coarse with respect to climate. For instance, ARI Climate Zone 2 in
16、cludes the widely varying climates of Phoenix (1 125 heating degree-days HDD/4189 cooling degree-days CDD, 99% Design Temp = 37“F, 1% Design Temp = 108F) and Ft. Worth, Texas (2370 HDD/ 2568 CDD, 99% Design Temp =24“F, I% Design Temp = 98“F). Although it was never intended that the Region IV HSPF an
17、d SEER would be used to estimate energy use across climates, these values have indeed been used for these purposes within software and calculation procedures (for instance, see ACCA 19861). Particularly for heat pumps, this can lead to erroneous conclusions on their relative merit as compared with n
18、on-heat pump alternatives across climates. Given these limitations, it is highly desirable to have some means to interpret the seasonal performance of heat pumps and air conditioners across locations. Beyond the climatic variation, there are other reasons that typical performance may not reach that
19、suggested by the AM standard. Within the 2 10/240 procedures, a correction factor, C, with a value of 0.77, is used to reduce the heating loads on the heat pump. The justification for the C-factor is that it more closely matches measured building loads when used with degree-day or bin weather data b
20、ased on a 65F setpoint (Harris et al. 1965). The reason the building loads are lower is due to heat gains, such as solar gains and internal gains, which cause a balance point lower than 65F. A better method to account for this is to use a lower balance point rather than a multiplier since the effect
21、 at varying outdoor temperatures can be very different from the C= 0.77 default. For instance, tests performed by the Electric Power Research Institute in the late 1970s found that C could vary from 1.2 to 0.4 in actual resi- dences (EPRI 1980). As a result, the method of estimation has fallen out o
22、f favor, as evidenced by its disappearance from the ASHRAE Handbook-Fundamentals after 1985. Another problem associated with the AR1 procedure is that it implicitly assumes a 65F interior heating setpoint by using bin data at 65F along with C to reduce those loads. In our analysis, we desired to use
23、 the more commonly preferred 68F interior heating setpoint. This will have the effect of increasing the building load, which, in turn, will tend to reduce the simulated HSPF as compared with the assumption of 65“F, as in ARI 210/240. Backup resistance use will also increase. At an outdoor temperatur
24、e of 45“F, this difference could impose a 30% increase in loads after internal and solar gains are taken into account. I. Similarly, the mild climate of Los Angeles falls into the same clas- sification as Atlanta, Georgia, and St. louis, Missouri, whereas each of these has widely varying weather. Co
25、oling Mode Operation and SEER The seasonal energy efficiency ratio (SEER) rating for central air conditioners was adopted in 1979 after years of development. Last modified in 1994, SEER is a national metric that does not account for regional differences in summer climate. In addition, the SEER ratin
26、g deemphasizes high-temperature performance (EER,). Indeed, for single- speed equipment, the test is entirely based on performance at 82“F, so designs can be (and often are) optimized for moderate temperatures. Even for modulating equipment, the weight assigned to high-temperature performance is ver
27、y low. Since high-temperature performance can vary significantly among designs with the same SEER, and since high-temperature performance is a key determinant of utility peak loads, this limitation is of concern. In addition, the AR1 standard speci- fies static pressures that are about half as large
28、 as the averages from field studies (Proctor and Parker 2000). High static pres- sure means lower airflow across the evaporator coil, which leads to a relatively cold coil, which increases cooling energy use in dryer climates and lowers heat pump performance everywhere. Similarly, the AM procedure i
29、mplicitly assumes an 80F interior setpoint for cooling as opposed to the 78F setpoint assumed in this study. The more realistic but lower setpoint will also result in greater loads for the “operating efficiency“ calculations performed here and, hence, result in lower predicted cooling performance. T
30、hese limitations of the assumptions used within the SEER procedure have been pointed out by others (Kavanaugh 2002). When used in cooling mode, the performance of a heat pump is greater due to the smaller temperature differences between indoor and outdoor conditions and the fact that heat is being e
31、xtracted from the interior of the home rather than from a very cold exterior condition. Thus, typical cooling system SEERs are on the order of 10-17 Btu/Wh for current generation equipment. For the analysis presented here, systems with SEERs equal to or greater than 13.5 Btu/Wh are assumed to have e
32、lectronically commutated motors (ECM) in use within the indoor blower units. Likewise, the analysis assumes that ECMs are utilized where HSPF is equal to or greater than 8.5. Empirical Tests of Heat Pump Performance As heat pump technology emerged in the early 197Os, a number of measurements were ma
33、de on heating system performance. Many laboratory studies were performed under steady-state conditions to evaluate the impacts of defrost cycles, crankcase heat, and other influences (e.g., Parken et al. 1977; Rettberg 1980).2 It has been long known that even with a constant thermostat setting, when
34、 building loads exceed the declining capacity of the heat pump, the difference must bc made up with resistance heat. This will impact overall effi- ciency (Reedy and Daniels 1992). 2. Percentage increases to seasonal heat pump energy consumption associated with evaporator defrost vary from 4% to 10%
35、 with 8% being typical (Gross et al. 1978; Baxter and Moyers 1985). ASHRAE Transactions: Research 1 79 Many ofthe existing early studies do show that heat pump performance was often lower than would be expected by the AM procedures, In a study for a Louisiana utility, Orth et al. (1 976) performed a
36、lternate day resistance heating measure- ments on two 1967 vintage heat pumps and found seasonal coefficient ofperformance (SCOP) measurements of 1.75 and 1.78 for the systems, as compared with the manufacturers SCOP rating of 2.25. In the colder climate of New Jersey, Nicolich (1 977) estimated the
37、 SCOP of a single heat pump to be 1.65 based on pre- and post-measurements. In the much colder climate of Ontario, Canada, 40 heat pumps were moni- tored in detail, showing average SCOPs of 1.43 over the heat- ing season in 1975-1977. Similarly, a large study (Groff et al. 1978) showed average seaso
38、nal COP values of 1.61 to 1.2 in the Boston and Minneapolis climates, respectively. However, even in moderate climates, performance may be less than anticipated. Four residences in Albuquerque, New Mexico, that had the heating system performance of their heat pumps evaluated through alternate day re
39、sistance heat opera- tion, showed SCOPs averaging only 1.39 (1.42,l. 10,1.64, and 1.4 1) as opposed to the HSPF calculation, which indicated a SCOP of 1.85 (Wildin et al. 1978). Significantly, the study estimated that homeowner operation of thermostats led to lower than expected savings. Another stu
40、dy in Knoxville, Tennessee, ofhvo heat pumps (Baxter 198 1) yielded measured SCOPs of 1.58 and 1.99, respectively, as compared with labeled SCOPs of 1.99 and 2.6 1. Admittedly, the above studies were on early generation heat pumps rather than more modem equipment. However, the relevance of these old
41、er investigations is underscored by more recent data suggesting similar trends. Recent submetered data from over 160 Florida homes show that operational heat pump performance is adversely impacted by thermostat setback (Bouchelle et al. 2000). This same phenomenon had been observed in earlier resear
42、ch on monitored heat pump perfor- mance (Bullock 1978). Since the utility experiences its annual system peak during Floridas few cold mornings, the perfor- mance of heat pump systems is important to controlling demand. While the mild conditions prevailing should allow heat pumps to operate under fav
43、orable conditions, data anal- ysis revealed a large impact from auxiliary electric resistance strip heat on site-achieved heat pump efficiency. Households practicing temperature setback followed by a morning setup (about 25% of total) showed large amounts of strip heat during morning operation, whic
44、h reduced overall coefficients of performance (COP).3 Thus, the measured annual space heat consumption of the 9 1 heat pump systems was 1,038 kWh, as compared to 1,292 kWh for the 57 homes in the electric resistance group. Correcting for differing floor areas, the group of homes with heat pumps show
45、ed a nonnal- ized energy use of 0.65 kWh/ft2 against 0.94 kWWft2 for the homes with electric resistance forced-air heating systems. The implied seasonal COP from these data is only 1.45 compared with the commonly claimed 2.0 (HSPF = 6.8). 3. Approximately 5% of audited households had a non-heat-pump
46、 thermostat so that such systems operated exclusively in strip heat mode. Other customers operated the thermostat into “emergency heat” mode, which exclusively uses strip heat. Although this impacted the utility savings of heat pumps, we did not simulate these performance-related real-world effects.
47、 Another contemporary source of empirical data on comparative heat pump performance comes from the Pacific Northwest in submetered data taken by the Bonneville Power Administration. As part of its Super Good Cents program, hundreds of homes had space heat submetered from 1987 to 199 1, with detailed
48、 audit information on the homes (Andrews et al. 1989; Echan 2000). Within these data, homes in the populous coastal region of the Pacific Northwest showed an average measured annual space heat of 7,841 kWh (3.63 kWh per square foot of floor area) for those with heat pumps (n = 85) against 8,953 kwh
49、(4.46 kWh per square foot) for those with forced air electric strip heat (n = 35). Although the savings produced by heat pumps was statistically significant, the implied coefficient of performance was only 1.23-well below the nameplate COPS of 1.99 or better! Although not evaluated here, previous monitoring and evaluation have shown that thermostat setback with morning setup can have very deleterious effects on air-source heat pump performance as the sudden increase in morning thermo- stat setup triggers the use of lower efficiency auxiliary resis- tance stri
