Experimental research and operation optimization of an air-source heat pump water heater
J.J. Guo,J.Y. Wu,R.Z. Wang,S. Li
DOI: https://doi.org/10.1016/j.apenergy.2011.04.012
IF: 11.2
2011-01-01
Applied Energy
Abstract:In order to optimize design and operation strategy of air-source heat pump water heater, an experimental set-up and simulation model were constructed. Also, a methodology of optimizing operation, which takes into account thermostatic and timing control patterns, was presented and applied. Experimental results indicated that the average COP ranged from 2.82 to 5.51 under typical conditions. The recommended outside area ratio of condenser coil to evaporator is 0.14−0.31 when the evaporator outside area is between 6.0 and 6.5 m 2 for this set-up. The optimal start-up time was between 12:00 and 14:00 if there was no electricity price difference between day and night, or it was near 22:00. The optimal setting water temperature should be adjusted according to the variation of seasonal ambient temperature. It was suggested that, based on this set-up, setting water temperature should be set higher than 46 °C in summer and 50 °C in other seasons. Keywords Air-source heat pump water heater Experimental research Optimization Nomenclature A heat exchange area, m 2 C specific heat, KJ/(kg K) COP coefficient of performance d diameter of copper tube, m f operation frequency of compressor, Hz f(t) domestic hot water demand profile, kg/s g(t) hot water flow rate from water tank, kg/s h specific enthalpy, KJ/kg j number order of time step k 1 , k 2 modifying factor of thermostatic expansion valve m ˙ mass flow rate of refrigerant, kg/s M mass, kg n polytropic index P pressure, Pa p pitch, mm Q ˙ heat transfer rate, kW r latent heat, kJ/kg s revolutional slip t time, s T temperature, °C U heat transfer coefficient, kW/(m 2 K) v specific volume, m 3 /kg V volume, L V d compressor displacement volume rate, cm 3 /rev W total input power, kW W(t) electrical power consumption of compressor and fan, kW Subscripts a ambient; air cmp compressor cond condensation or condenser consume consumed water eva evaporation or evaporator fin fin i inside; inlet l refrigerant liquid load hot water load lon longitudinal tube o outside p constant pressure r refrigerant s sensible heat transfer set setting water temperature sc subcooled sh superheated tap tap water tp two-phase tra transverse tube tube heat transfer tube v volumetric; vapor w water wi initial water Superscripts Δ t testing time step between the initial and end time 0 testing initial time Greek letter δ fin thickness of evaporator, mm η efficiency α heat transfer coefficient, kW/(m 2 K) λ thermal conductivity, kW/(m K) ξ dehumidifying coefficient ρ density, kg/m 3 η fin general fin efficiency ζ delivery coefficient Δ τ duration of one time step in numerical calculation 1 Introduction The proportion of water-heating in building energy consumption is about one-third. Nowadays, with more concern is taken to building energy-saving, considerable importance has been placed on the energy efficiency of water heater in buildings [1,2] . The air-source heat pump water heater (ASHPWH) system, a water heating device with about one-third energy consumption of the conventional electric resistance water heater, has attracted more and more interests for its high efficiency, energy-saving and environmental benign recently [3,4] . The thermal performance of the ASHPWH system is influenced by refrigerants, system structure, ambient temperature, water temperature, etc [5–12] . And many achievements have been obtained by the researchers. Mei et al. [13] investigated the difference of R22 and R407C in the practical ASHPWH system. Afterward, they studied the effect of different style natural convection immersed condensers to the heat pump water heater [14] . Morrison et al. [15] presented an ASHPWH seasonal performance assessment method based on experimental results. Their research indicated the effect of ambient temperature (surrounding air temperature) on the system performance is more significantly than the effect of initial water temperature during the heating process. Rankin et al. [16] presented a study about demand side management for commercial building using an inline air-source heat pump water heater methodology. Kim et al. [17] suggested that the tank size should be optimized through simulation. In Japan, the scholars [18] had studied the influence of ambient temperatures to the whole system, and pointed out that operation strategy of ASHPWH system is greatly important and should be considered in their subsequent work. However, due to the variation of the ambience and heating load, ASHPWH systems are only optimal under certain working conditions if the structure is fixed. Besides, if suitable operation strategy is established, such as selecting correct start-up time or setting water temperature for the system, satisfactory thermal and economic performance can be also obtained [19,20] . Rankin et al. [16] solved the peak demand of national electricity supply grid by controlling the operation of inline air-source heat pump water heater and auxiliary heater simultaneously. Huang et al. [21] presented a fast response heat pump water heater by controlling different water temperatures of two tanks, and the key of its operation strategy is to setting different priority for two tanks. As discussed above, the operation strategy is greatly subject to the consumer’s control patterns, in practical application, thermostatic and timing control patterns are two commonly control patterns adopted by the consumers. Thus, the operation strategy optimization should be considered according to different control patterns. Whereas the potential of such operation strategy optimization is far from being developed, there is lack of information about the operation strategy optimization of ASHPWH systems, especially in the residential buildings. In this paper, structural and operation strategy optimization of an experimental set-up of air-source heat pump water heater (ASHPWH) system were investigated. Based on a simulation model, the relationship between system thermal performance and outside area ratio of condenser coil to evaporator was discussed. Additionally, a methodology for the operation strategy optimization of the system was introduced and applied, and the optimal start-up time and setting water temperature were obtained based on timing and thermostatic control patterns, respectively. 2 Description of system and experiment set-up 2.1 Working process of the ASHPWH system The system is composed of outdoor and indoor units, as shown in Fig. 1 . The outdoor unit, such as compressor, filter–drier, thermostatic expansion valve (TXV) and the evaporator, is mounted in the environmental chamber. The other parts of the ASHPWH system, for example water tank and condenser, is installed in the ambient out of the environmental chamber. The condenser is immersed in the hot water tank. The air temperature in the environmental chamber can be adjusted from 5 °C to 40 °C. The working process undergone by the refrigerant during a cycle can be represented by an idealized heat pump cycle as shown in Fig. 2 . Here, 1–2, 2–3, 3–4 and 4–1 represent compression, condensation, throttling and evaporation processes, respectively. After the throttling process, the refrigerant in the evaporator will absorb the heat from the ambient air, then after the compression process, the heat will be released into the water tank. As the water tank getting warmer, the working process cycle will be updated accordingly, such as 3′–4′–1′–2′–3′. 2.2 Experimental facility and procedures An experimental set-up of ASHPWH system was developed in the lab. Fig. 3 and Table 1 show the details of indoor and outdoor units. In this experiment, the evaporator, TXV and the compressor were mounted in an environmental chamber, as shown in Fig. 3 a and b. The rated input power of the rotary-type hermetic compressor was 0.93 kW. To avoid the overload, an overheated protector and low-high pressure cut-off switches were connected to the compressor. The evaporator was a fin-tube heat exchanger. The condenser was made up of a smooth copper helical coiled tube (Ø9.90 * 0.75), which was immersed in the water tank. The curvature radius of the coil was 0.16 m. The filter–drier and TXV were installed downstream the condenser. The whole system was controlled by a micro-controller. Before the experiment, the environmental chamber was adjusted at certain value (it represents ambient temperature), and the tap water was charged into the 150 L water tank. Then, the water was heated after the ASHPWH system started. If the water temperature in the tank reaches to the setting point, the system is stopped by the micro-controller. In the process of hot water consuming, the hot water was drawn from the top of the tank, and cold tap water was recharged from the bottom of the tank. The main aim of the experimental work is to test the system thermodynamic performance, such as the COP and the average COP. (1) COP = C p , w M w ( T w Δ t - T w 0 ) ∫ 0 Δ t W ( t ) dt As shown in Eq. (1) , the COP is the ratio of heat transfer from the condenser to water to the electric power consumption of compressor and fan. The tank water temperature (i.e. T w Δ t or T w 0 ) is represented by the average value of the three platinum resistance thermometers (i.e. Pt100) set at the bottom, central and top of the tank, as shown in Fig. 1 . 0 and Δ t represent the testing initial time and time step between the initial and end time, in which there is 3 °C increment in water temperature. The heating capacity is the time average heat transfer from the condenser to water within 3 °C increment in the water tank. ∫ 0 Δ t W ( t ) dt represents the electrical power consumption of compressor and fan within time step measured by power meter unit, kJ. Eq. (1) represents COP at water temperature ( T w Δ t + T w 0 ) 2 . When the water in the tank is heated from certain initial temperature (i.e. T wi ) to the setting point (i.e. T set ), the average COP during the whole process is (2) COP ave = C p , w M w ( T set - T wi ) ∫ 0 t W ( t ) dt where t represents the duration of the whole process. ∫ 0 t W ( t ) dt represents the electrical power consumption of compressor and fan measured by power meter unit during the whole process. 2.3 Error analysis The power meter unit (with grade 0.5 accuracy, uncertainty about ±10 W) was used to measure the electrical power consumption of the compressor and fan automatically. The three platinum resistance thermometers (i.e. Pt100, with grade A, uncertainty of ±0.2 °C) inserted along the axis of the tank, were set at the bottom, central and top of the tank. All the data were collected by a data acquisition system and stored in a computer. The interval of data collection can be set fifteen seconds, and the error of the data acquisition system can be ignored. Therefore, the error analysis of the experimental results on the basis of the uncertainties in the measurements was performed using the Kline and McClintock relationship [22] , and the relative error of testing instrument was 9.4% for COP, 0.8% for average COP, respectively. 3 Mathematical model and simulation process In order to predict the operation performance of ASHPWH system and find the influence of various parameters on its operation, a mathematical model has been formulated in this paper. Some physical assumptions employed for the present model are: (1) Compression of refrigerant vapor is assumed to follow a polytropic process. (2) Pressure drop is negligible in evaporator, condenser as well as pipes. (3) Expansion of refrigerant liquid is considered to be isenthalpic. 3.1 Compressor model Since the compression of refrigerant vapor is assumed to be a polytropic process, the energy equation for the compressor can be expressed as: (3) m ˙ r ( h 2 - h 1 ) = m ˙ r · P eva · v 1 · n n - 1 P cond P eva n - 1 n - 1 where n is the polytropic index. The mass of refrigerant ( m ˙ r ) pumped by the compressor is calculated by [23] , (4) m ˙ r = m ˙ cmp = f ( 1 - s ) V d η v v 1 where V d is the displacement volume, s is the revolutional slip and η v is the volumetric efficiency. 3.2 Evaporator model The physical components and working environments of the evaporator in the ASHPWH system, though the working condition in summer is formidable, is similar to the normal air-source unitary heat pump. So the evaporator model of the system can be presented as follow. The energy equation for the evaporator can be expressed as (5) Q ˙ eva = m ˙ r ( h 1 - h 4 ) = U tp , eva A i , eva - C r , sh , eva m ˙ r U sh , eva · ln T a - T 1 T a - T eva · ( T a - T eva ) + C r , sh , eva m ˙ r ( T 1 - T eva ) where U tp,eva and U sh,eva denote the heat transfer coefficient of the two-phase section and superheated section in the evaporator, respectively, which are expressed as: (6) U tp , eva = 1 α tp , eva + d i , eva 2 λ tube ln d o , eva d i , eva + 1 α o , s , eva · ξ · η fin · A i , eva A o , eva - 1 (7) U sh , eva = 1 α sh , eva + d i , eva 2 λ tube ln d o , eva d i , eva + 1 α o , s , eva · ξ · η fin · A i , eva A o , eva - 1 where α o , s , eva is sensible heat transfer coefficient at the outside of evaporator, which is developed by Wang and Du [24] . α tp , eva and α sh , eva represent the heat transfer coefficient of the two-phase section and superheated section at the inside of evaporator, which are expressed in [25] . 3.3 Condenser model The condenser energy balance equation can be expressed as: (8) Q ˙ cond = m ˙ r ( h 2 - h 3 ) = C r , sh , cond m ˙ r ( T 2 - T cond ) + m ˙ r r cond + U sc , cond A i , cond - C r , sh , cond m ˙ r U sh , cond · ln T cond - T w T 2 - T w - m ˙ r r cond U tp , cond ( T cond - T w ) · T cond - T 3 ln T cond - T w T 3 - T w where Q ˙ cond is the condensing heat transfer rate, U tp,cond , U sh,cond , and U sc,cond denote the heat transfer coefficient of the two-phase section, superheated section and subcooled section in the condenser, respectively, which are expressed as: (9) U tp , cond = 1 α tp , cond + d i , cond 2 λ tube ln d o , cond d i , cond + 1 α w · A i , cond A o , cond - 1 (10) U sh , cond = 1 α sh , cond + d i , cond 2 λ tube ln d o , cond d i , cond + 1 α w · A i , cond A o , cond - 1 (11) U sc , cond = 1 α sc , cond + d i , cond 2 λ tube ln d o , cond d i , cond + 1 α w · A i , cond A o , cond - 1 where α tp , cond , α sh , cond and α sc , cond represent the heat transfer coefficient of the two-phase section, superheated section and subcooled section at the inside of condenser, and α w is the heat transfer coefficient in the water side. The four heat transfer coefficients are suggested as [26] . 3.4 TXV model Since the expansion of refrigerant is assumed to be isenthalpic, the TXV energy balance equation is: (12) h 3 = h 4 The mass of refrigerant ( m ˙ r ) through the TXV can be calculated by [27] (13) m ˙ r = m ˙ TXV = ( k 1 + k 2 · T eva ) · ρ l ( P cond - P eva ) where k 1 and k 2 were found to be equal to 3.578 × 10 −5 m 2 and 2.442 × 10 −7 m 2 /K for the valve considered in present study. 3.5 Water tank model The water is insulated and the helical coiled tube type condenser is immersed in the water tank, therefore the heat loss from the water tank to ambient is negligible. The variation of heat storage in the water tank is the combined result of condensing heat release and hot water load, see Hawlader et al. [23] for a detailed account of the water tank model. The energy balance equation can be obtained by (14) d dt ( C p , w M w T w ) = Q ˙ cond - Q ˙ load Before the hot water is delivered to the end user, the hot water flow from water tank and cold supply water (tap water) would be mixed. The energy conservation equation is expressed as (15) C p , w · g ( t ) · T w + C p , w · [ f ( t ) - g ( t ) ] · T tap = C p , w · f ( t ) · T consume where g ( t ) is hot water flow rate from water tank, which depends on the domestic hot water demand profile f ( t ). Where T consume is the temperature of consumed hot water, and T tap is the tap water temperature. 3.6 Performance index The COP of the system performance (16) COP = ∑ j = n 1 n 2 Q ˙ cond ( j ) × Δ τ ∑ j = n 1 n 2 W ( j ) × Δ τ where Δ τ is the duration of one time step in numerical calculation. n 1 and n 2 represent the n 1 th – n 2 th time step. Q ˙ cond ( j ) and W ( j ) are condensing heat transfer rate and the electrical power consumption of compressor and fan at the j th time step. In order to verify the calculated results with the experimental data, 3 °C water temperature increment is chosen from the n 1 th to n 2 th time step. The water temperature at the n 1 th and n 2 th time step are T w n 1 and T w n 2 , respectively. Hence, the calculated COP can be regarded as the COP at water temperature ( T w n 1 + T w n 2 ) 2 . The average COP during the whole heating process, (17) COP ave = ∑ j = 0 n Q ˙ cond ( j ) × Δ τ ∑ j = 0 n W ( j ) × Δ τ where 0 and n are the first and last time steps of the whole heating process, respectively. 3.7 Computing procedure Based on the mathematical model developed for this ASHPWH system, a simulation program was developed to estimate the thermal performance of the system. The input data for simulation were the structural parameters, meteorological data and the initial water temperature in water tank. The heating capacity, electrical consumption, COP of the system and all the state points data (such as, temperature, pressure, specific volume, specific enthalpy and the mass flow rate), were calculated as output data. The properties of the chosen refrigerant R22 were from REFPROP (version 6.01) developed by NIST (National Institute of Standard and Technology). 4 Results and discussion 4.1 Experimental results and model verification In order to learn the performance of the ASHPWH system, a series of experiments were carried out in the environmental chamber under different ambient conditions based on the meteorological data of Shanghai, as shown in Figs. 4–6 and Table 2 . In the experiments, the ambient temperature was set constant. After the ASHPWH system starts, the 150 L water in the tank was heated from initial temperature to 55 °C, and then the system will be stopped by the micro-controller. Fig. 4 shows the variation of condensing and evaporating pressure when the water tank was heated from 15° C to 55 °C in winter (i.e. T a = 5 °C). The simulated results agree well with the experimental results. For the condensing pressure, the largest deviation between the experimental and simulated results was less than 0.5Bar. For the evaporating pressure, the largest deviation between the experimental and simulated results was less than 0.3Bar. Fig. 5 shows the comparison of heating capacity variation when water tank temperature was from 16.5 °C to 53.5 °C during operation. For winter case (i.e. T a = 5 °C), the experimental heating capacity decreased gradually from 2.46 kW to 1.99 kW. While for summer case (i.e. T a = 35 °C), it increased from 4.18 kW to 4.57 kW firstly, and then decreased to 4.05 kW. For winter case, with the increase of water temperature and condensing temperature, the enthalpy difference between the inlet and outlet of the condenser was decreasing ( Fig. 2 ), which led to lower heating capacity. For summer case, when condensing temperature was lower than the ambient temperature, the pressure difference between the condenser and evaporator was too small, which led to lower heating capacity. With the increase of water temperature, the pressure difference was increasing, and then yielded more heating capacity. When the condensing temperature increased to some value, the decrease of the enthalpy difference between the inlet and outlet of the condenser dominated, which caused lower heating capacity. Also, the figure shows that the simulation results were in good agreement with the experimental data with maximum deviation of 9.8%. This attributed to actual pressure drop and heat loss of the system. As described above, the heating capacity of the ASHPWH system in the experiments varied with the water temperature and ambient temperature, which led to the variation of COP. Fig. 6 shows the experimental and simulated COP variation with the water temperature from 16.5 °C to 53.5 °C during operation. The COP decreased with increasing water temperature for both summer and winter cases. The maximum deviation between experimental and simulated results at the same water temperature was less than 10%. Table 2 shows the comparison of the experimental and simulated results of system average COP during the whole heating process, under typical working conditions. The experimental results show that the average COP during the whole heating process was 5.51 in summer ( T a = 35 °C, T wi = 10 °C) and 2.82 in winter ( T a = 5 °C, T wi = 25 °C), which was the peak and valley of average COP in the whole year, respectively. The experimental average COP was lower than the simulated average COP. This was due to the pressure drop and heat loss during actual operation, which in turn caused more compressor work and lower efficiency. The maximum deviation between the experimental and the simulated results was less than 10%. The cases study showed good agreement between the experimental data and the simulation results. Hence, the model was reliable for further analysis. 4.2 Optimization analysis based on structural parameters and control patterns In order to have a more understanding on the factors affecting the performance of the system and to be able to optimize its performance, more theoretical work was done. Based on the simulation model, the effect of structure parameters and two different control patterns were analyzed in the following section. 4.2.1 Optimization analysis based on structural parameters As discussed above, such a system is most suitable for residential buildings in mild weather and domestic hot water is needed all year round. To ensure well performance of the system, it is necessary to learn the effect of variation of the evaporator and condenser coil area on the system performance. As shown in Fig. 7 , the 150 L water in the tank was heated from 15 °C to 55 °C. It could be seen that, initially, the average COP increased with the increase of the evaporator and condenser coil outside area obviously. However, if the evaporator outside area increases beyond a certain value, the slope of average COP trends to be flat. This indicated that the effect of evaporator and condenser coil outside area on average COP became less sensitive, based on this kind of heat exchanger. It also indicated that it was unnecessary to enlarge heat exchange area excessively when a relatively high average COP has been obtained. Considering the actual finite assembling room and material saving, satisfactory performance can be obtained when the evaporator outside area is between 6.0 and 6.5 m 2 , and the recommended outside area ratio of condenser coil to evaporator is between 0.14 and 0.31. 4.2.2 Relationship between thermostatic and timing control patterns In order to optimize the operation strategy, selecting correct control pattern is a premise. In practical application, thermostatic and timing control patterns are two common control patterns. For the ASHPWH system, the thermostatic control pattern is very widely used, in this pattern, water temperature of the ASHPWH system is set at T set (setting water temperature). If the water temperature is lower than T set , the ASHPWH system will turn on automatically, and the ASHPWH system turns off until water temperature reaches to T set [11] . The advantage of this pattern is simple and convenient. While, the disadvantage of this pattern is that water tank is almost kept at T set , as a result, the ASHPWH system always operates when condensing pressure is high, which lowers the system average COP. For the timing control pattern, when the ASHPWH system turns on, water is heated from initial water temperature to T set and then turns off before domestic hot water withdrawal. After the withdrawal, the water temperature in the tank is lower than T set until next start-up of ASHPWH system. As to the start-up times, it depends on the practical total hot water consumption volumes of consumers. Since water in the tank is always heated from initial water temperature to T set , but not almost kept at T set like in the thermostatic control pattern, thus, it is more energy-saving because higher system COP can be obtained at relative lower water temperature [28] . The timing control pattern is suitable in some special sites (such as the consumer has the fixed domestic hot water consumption habit, the withdrawal frequency is low and interval between twice withdrawals is long). Therefore, the optimization operation strategy should be based on different control patterns, and a methodology of optimization analysis based on these two control patterns will be depicted in the subsequent contents. 4.2.3 Optimization analysis based on timing control pattern In this pattern, the start-up time determines the thermal and economic performance of the ASHPWH system. From energy-saving concern, higher COP can be obtained with the higher ambient temperature during the heating process [15] . So it seems that the optimal start-up time should be when the ambient temperature is higher. While, subject to the solar radiation, the diurnal ambient temperature is not always kept constant during the heating process, such as in Shanghai, China. Additionally, the policy of peak and valley electric power price was carried out in some cities, which impelled the electric consumption at night. However, the COP of the ASHPWH system drops for the lower ambient temperature at night. Energy-saving and money-saving are not always coincident, so that the two factors (fluctuation of ambient temperature and policy of peak and valley electric power price) should be synthetic considered in timing operation pattern. In order to illustrate the effect of two factors to the whole ASHPWH system, the case of Shanghai was discussed as following. Fig. 8 shows the ambient temperature of a typical day in Shanghai, the Stages A and B represent peak and valley time of a whole day with different electric power prices 0.088 US$/kWh and 0.044 US$/kWh. Figs. 9–11 show the energy consumption, electric fees and average COP during the whole heating process when 150 L water of the ASHPWH system was heated from 15 °C to 55 °C from certain time in a typical day. Fig. 9 shows the case in summer. It could be seen that the valley energy consumption, electric fees and highest average COP during Stage A were 1.3 kWh, 0.113 US$ and 5.43 respectively when water tank was heated from 14:00 for the highest ambient temperature, and the system got the highest performance, while the valley electric fees of the whole day was 0.065 US$ at 22:00 for the valley electric power price. For the same reason, the valley electric fees were 0.079 US$ and 0.094 US$ respectively, as shown in Figs. 10 and 11 , both at 22:00. It was suggested that the optimal start-up time was near 22:00 for the lowest electric power fees. However, if the electric power price of the day and night was equal, the optimal start-up time should be chosen between 12:00 and 14:00. It could be derived that the operation strategy is affected by the price difference of peak and valley electric power price. The smaller the price difference of peak and valley, the more economic the operation is in the daytime. The curves of fee in Figs. 9–11 also show that the electric power fee was very near when the start-up time was from 22:00 to the next 4:00 in summer and transitional season, and from 22:00 to the next 2:00 in winter. According to the current policy of peak and valley electric power price, the start-up time of ASHPWH system is better to selected as close as possible to these time ranges. As to the start-up times and total energy consumption, it depends on the practical hot water consumption volume in some special sites. 4.2.4 Optimization analysis based on thermostatic control pattern In the thermostatic control pattern, the water temperature in the tank is controlled according to T set (setting water temperature). Usually, T set is set to be a fixed value, such as the ASHPWH system with R22, it is always set to 55 °C for hot water consumption habit concern. Whereas, from the viewpoint of energy-saving, decreasing the T set will improve the COP of the whole system. Additionally, a too low T set will lead to the hot water shortage if the hot water consumption volume is large. Therefore, there must be a compromise between T set and consumed volume of domestic hot water. However, the consumed volume of domestic hot water as well as the consumption pattern with time varies from household to household [29] , it is therefore difficult to define general domestic hot water demand profiles [30–32] . Considering the diurnal and seasonal variation of domestic hot water consumption, Roman Spur et al. [33] presented a typical scheme which could contain wide scale of domestic hot water usages in household, as shown in Fig. 12 . The consumption volume of domestic hot water was non-homogeneous during the typical day, and the domestic hot water demand profile was classified into three levels: heavy, medium and light [33] . Similarly, domestic hot water demand of a common family is usually the highest in winter, followed by transitional season and summer, which also indicates three levels of hot water consumption. Thus, the three levels domestic hot water demand profiles presented by Spur were assumed the daily hot water consumption of a family in winter, transitional season and summer in this study. In order to learn the effect of T set on the ASHPWH system performance, so as to establish suitable operation strategy, a case of a family in Shanghai was investigated. The ambient temperature was based on the description in Fig. 8 . The (a), (b) and (c) in Fig. 12 were assumed a typical day hot water demand profile f ( t ) in summer, transitional season and winter, respectively. The heating performance of the ASHPWH system was shown in Figs. 13–15 , which represent the typical seasonal performance. An evaluating indicator called delivery coefficient (i.e. ζ ) was introduced, and it was defined as the ratio of actual heat the user consumed hot water to the domestic hot water demand during a day time, as below (18) ζ = ∫ 0 86400 Q ˙ cond dt C p , w V consume ρ w ( T consume - T tap ) where the V consume represents 100 L, 180 L and 320 L when it is in summer, transitional seasons and winter, respectively. The temperature of tap water and consumed domestic hot water are 15 °C and 45 °C, respectively, in the simulation. Fig. 13 shows the influence of T set on the electric consumption of the system. With the decrease of T set from 55 °C to 46 °C, the electric consumption of the system during a typical day in summer, transitional season and winter decreased by 20.1%, 20.6% and 22.8%, respectively. It seems that a lower T set should be a better choice for improving the system performance. Nevertheless, such a strategy may lead to the hot water supply trouble, as discussed before, the hot water shortage. Such an effect was depicted in Fig. 14 detailedly. Fig. 14 shows the user consumed water temperature under different T set during a typical day in different seasons. The region size with temperature below 45 °C (the required temperature of domestic hot water) indicated the shortage extent of hot water. It can be seen that hot water shortage extent was quite different with different T set and different seasons. The hot water shortage extent can be characterized by two parameters: the shortage time and the temperature valley value of user consumed water, which somewhat like the bottom and height of a triangle. With a same T set , the shortage time was shortest in summer and longest in winter, and the temperature valley value was highest in summer and lowest in winter. The case in transitional season was in the middle of summer and winter. With the increase of T set , the hot water shortage will be relieved. In summer, the hot water load is relatively light, thus even a very low T set can achieve the goal of hot water supply. As shown in Fig. 14 a, even when the T set was set as low as 46 °C, which was very close to the required temperature 45 °C, the shortage time in the whole day was less than fifteen minutes and the temperature valley value was higher than 40 °C. In this case, the hot water shortage extent was less than 1%, which can be seen in Fig. 15 (delivery coefficient ζ > 99%). Such a light shortage can be neglected and is acceptable for the user. Therefore, the T set could be set as low as 46 °C in summer. In winter, the hot water load is relatively high, thus a low T set might be unable to afford the hot water requirement. As shown in Fig. 14 c, if the T set was set as 46 °C (same as the case in summer), the hot water shortage will be very large. In this case, the shortage time lasted about 1.5 h and the temperature valley value was as low as 33 °C. Such a hot water shortage was too heavy and unacceptable. Therefore, a higher T set was required. As shown in the figure, when the T set was set as 52 °C, the hot water shortage will be eliminated. According to above discussion, the T set influenced both the energy performance and the hot water supplement performance of the system, and these two effects contradicted each other. The T set can not be too high and too low, and there will be an optimum value of T set in every season for the ASHPWH system. Fig. 15 shows the influence of T set on the daily average COP and delivery coefficient during a whole day of the system. According to the definition, the delivery coefficient reflects the hot water load satisfaction degree. If there is no hot water shortage during the whole day, the delivery coefficient will be 100%, otherwise, the delivery coefficient will be smaller than 100%. As shown in Fig. 15 , the daily average COP increased with the decrease of T set for the lower water temperature and condensing temperature. While the delivery coefficient declined with the decrease of T set . Considering that the delivery coefficient should be over 99%, which ensures the shortage time for the user to wait less than 20 min, the T set should be set above 46 °C, 49 °C and 51 °C, in summer, transitional season and winter, respectively. These values were the optimum T set of the system for different seasons, and it was obvious that the optimum setting water temperature should be adjusted according to the variation of seasonal ambient temperature. If the T set was lower than these values, the hot water supplement can not meet the hot water requirement. In these cases, it was not suitable to evaluate the system performance by the daily average COP. Under the optimum conditions, the daily average COP of the system were 4.3, 3.2 and 2.6, for the summer, transitional season and winter, respectively. 5 Conclusions An experimental set-up of air-source heat pump water heater (ASHPWH) was constructed and tested. Based on a simulation model, the optimization analysis of structural parameters was employed. Also, a methodology of optimizing operation, which takes into account thermostatic and timing control patterns, was presented and applied. The detailed conclusions are as follow: (1) The experimental results of this experimental set-up indicated that the average COP of the system under typical meteorological conditions of Shanghai ranged from 2.82 ( T a = 5 °C, T wi = 25 °C) to 5.51 ( T a = 35 °C, T wi = 10 °C) when water was heated to 55 °C. (2) The effect of evaporator and condenser coil outside area on average COP becomes less sensitive if outside area of condenser coil and evaporator increases beyond certain value. The recommended outside area ratio of condenser coil to evaporator is between 0.14 and 0.31 when the evaporator outside area is between 6.0 and 6.5 m 2 for this set-up. (3) As for the timing control pattern, the optimal start-up time was affected by fluctuation of ambient temperature and the price difference of peak and valley electric power. The smaller the price difference of peak and valley, the more economic the operation is in the daytime. The analysis about the typical day in Shanghai suggested that the optimal start-up time was between 12:00 and 14:00 if there was no electricity price difference between day and night, or it was near 22:00 for the lowest electric power fees. As to start-up times, it depends on the practical hot water consumption volume of consumers in some special sites. (4) Based on the thermostatic control pattern, an evaluating indicator delivery coefficient and the optimum setting water temperature were introduced. The optimum setting water temperature should be adjusted according to the variation of seasonal ambient temperature. It was suggested that, based on this system, setting water temperature should be set higher than 46 °C in summer and 50 °C in other seasons. Acknowledgments This research is supported by Hi-Tech Research and Development Program of China ( Grant No. 2007AA05Z220 ). Authors are particularly grateful to Dr. Qiang Ma for his technical support and to Mr. Yuxiong Xu and Yunkang Sun for their help in the experimental work. References [1] B. Ndoye M. Sarr Analysis of domestic hot water energy consumption in large buildings under standard conditions in Senegal Build Environ 43 2008 1216 1224 [2] Merih Aydinalp V.I. Ugursal S. Fung Alan Modeling of the space and domestic hot-water heating energy-consumption in the residential sector using neural networks Appl Energy 79 2 2004 159 178 [3] K.J. Chua S.K. Chou W.M. Yang Advances in heat pump systems: a review Appl Energy 87 2010 3611 3624 [4] T.T. Chow G. Pei K.F. Fong Z. Lin A.L.S. Chan M. He Modeling and application of direct-expansion solar-assisted heat pump for water heating in subtropical Hong Kong Appl Energy 87 2010 643 649 [5] P.G. Rousseau G.P. Greyvenstein Enhancing the impact of heat pump water heaters in the South African commercial sector Energy 25 2000 51 70 [6] L. Chen F. Meng F. Sun A novel configuration and performance for a two-stage thermoelectric heat pump system driven by a two-stage thermoelectric generator Proc IMechE Part A: J Power Energy 223 4 2009 329 339 [7] H. Yang P. Cui Z. Fang Vertical-borehole ground-coupled heat pumps: a review of models and systems Appl Energy 87 2010 16 27 [8] A. Nuntaphan C. Chansena T. Kiatsiriroat Performance analysis of solar water heater combined with heat pump using refrigerant mixture Appl Energy 86 2009 748 756 [9] K. Comakli F. Simsek O. Comakli B. Sahin Determination of optimum working conditions R22 and R404A refrigerant mixtures in heat-pumps using Taguchi method Appl Energy 86 2009 2451 2458 [10] Groll EA. Review of recent research on the use of CO 2 for air conditioning and refrigeration. Clima 2000, Napoli, Italy; 2001. [11] J. Zhang R.Z. Wang J.Y. Wu System optimization and experimental research on air source heat pump water heater Appl Therm Eng 27 2007 1029 1035 [12] G. Bourke P. Bansal Energy consumption modeling of air source electric heat pump water heaters Appl Therm Eng 30 2010 1769 1774 [13] V.C. Mei R.E. Domitrovic W.H. Brewer F.C. Chen Experimental study of an R407C heat pump water heater ASHRAE Trans 107 part 1 2001 224 229 [14] V.C. Mei F.C. Chen R.E. Domitrovic J.K. Kilpatrick J.A. Carter A study of a natural convection immersed condenser heat pump water heater ASHRAE Trans 109 part2 2003 3 8 [15] G.L. Morrison T. Anderson M. Behnia Seasonal performance rating of heat pump water heaters Solar Energy 76 2004 147 152 [16] R. Rankin P.G. Rousseau M.V. Eldik Demand side management for commercial buildings using an inline heat pump water heating methodology Energy Convers Manage 45 2004 1553 1563 [17] M.S. Kim M.S. Kim J.D. Chung Transient thermal behavior of a water heater system driven by a heat pump Int J Refrig 27 2004 415 421 [18] R. Yokoyama T. Shimizu K. Ito K. Takemura Influence of ambient temperatures on performance of a CO 2 heat pump water heating system Energy 32 2007 388 398 [19] A. Hepbasli Y. Kalinci A review of heat pump water heating systems Renew Sust Energy Rev 13 2009 1211 1229 [20] J. Stene Residential CO 2 heat pump system for combined space heating and hot water heating Int J Refrig 28 2005 1259 1265 [21] B.J. Huang J.H. Wang J.H. Wu P.E. Yang A fast response heat pump water heater using thermostat made from shape memory alloy Appl Therm Eng 29 2009 56 63 [22] J. Kline F.A. MoClintock Describing uncertainties in single sample experiments Mech Eng 34 1953 3 8 [23] M. Hawlader S.K. Chow M.Z. Ullah The performance of a solar assisted heat pump water heating system Appl Therm Eng 21 10 2001 1049 1065 [24] C.C. Wang Y.J. Du Airside performance of herringbone fin-and-tube heat exchangers in wet conditions Can J Chem Eng 77 6 1999 1225 1230 [25] V.V. Klimenko A generalized correlation for two-phase forced flow heat transfer Int J Heat Mass Transf 31 3 1988 541 552 [26] R.C. Jordan G.B. Priester Refrigeration and air conditioning 1948 Prentice-Hall New York [27] J.M. Jabardo Saiz W.G. Mamani M.R. Ianella Modeling and experimental evaluation of an automotive air conditioning system with a variable capacity compressor Int J Refrig 25 8 2002 1157 1172 [28] Y.W. Li R.Z. Wang J.Y. Wu Y.X. Xu Experimental performance analysis on a direct-expansion solar-assisted heat pump water heater Appl Therm Eng 27 2007 2858 2868 [29] Jordan U, Vajen K. Realistic domestic hot-water profiles in different time scales. IEA SHC Task 26: Solar Combisystems; 2001. [30] Mandate to CEN and CENELEC for the elaboration and adoption of measurement standards for household appliances: water-heaters, hot-water storage appliances and water-heating systems. Brussels: European Commission; 2002. [31] ENV 12977-2. Thermal-solar systems and components: custom-built systems, test methods. Brussels, CEN; 2001. [32] Service Water Heating. ASHRAE Handbook-HVAC System and Applications, USA; 1987. p. 17–20. [33] R. Spur D. Fiala D. Nevrala D. Probert Influence of the domestic hot-water daily draw-off profile on the performance of a hot-water store Appl Energy 83 7 2006 749 773