See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/311988058

Comparative research on different air conditioning systems for residential

buildings

Article  in  Frontiers of Architectural Research · December 2016

DOI: 10.1016/j.foar.2016.11.004

CITATIONS READS

5 829

3 authors:

Xin Zhou Da Yan

Tsinghua University Tsinghua University
31 PUBLICATIONS   365 CITATIONS    173 PUBLICATIONS   3,141 CITATIONS   

SEE PROFILE SEE PROFILE

Xing Shi
Southeast University (China)
58 PUBLICATIONS   522 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Modeling occupant behavior in buildings View project

2019 Researcher Links Workshop on Investigation of the Impact of Occupant Behavior on Building Performance in the UK and China View project

All content following this page was uploaded by Da Yan on 09 April 2018.

The user has requested enhancement of the downloaded file.

 1 Comparative research on different air-conditioning systems

2 for residential buildings

4 Xin ZHOU1,*, Da YAN2, Xing SHI1

1
7 School of Architecture, Southeast University, Nanjing, Jiangsu Province, 210096,

8 China

2
10 School of Architecture, Tsinghua University, Beijing 100084, China

11

*
12 Corresponding author. E-mail address: zhou-x06@seu.edu.cn (X. Zhou)

13

14

1
 1 Comparative research on different air-conditioning systems

2 for residential buildings

3 Abstract

4 There are generally two types of air-conditioning (AC) systems, namely, centralized

5 and decentralized AC systems. This study focuses on three actual engineering projects

6 of residential communities, where centralized AC systems are adopted. The

7 applicability of centralized AC systems in residential buildings is discussed and

8 analyzed. In addition, the key elements that lead to different building energy

9 consumptions and different system efficiencies between centralized and decentralized

10 AC systems in residential buildings are investigated. This study shows that in

11 residential buildings, at the point where the centralized feature of the system meets the

12 decentralized feature of users’ load, the problems of high energy consumption and low

13 energy efficiency could easily occur.

14 Keywords

15 Residential buildings, centralized AC system, decentralized AC system, comparative

16 research

17 1 Introduction

18 As a typical representation of a decentralized air-conditioning (AC) system, the split

19 type air conditioner has been used in a majority of residential buildings in China for a

20 long time. Recently, centralized AC systems have also appeared in residential

21 buildings, and they are approved and supported by specific government policies

2
 1 (Zhang et al. 2009). In some opinions, centralized AC systems reflect an advanced

2 and efficient way of energy usage. They consume less energy with better service;

3 therefore, the development of future indoor environment control in residential

4 buildings should take the centralized AC systems into consideration (Aste et al. 2013).

5 One of the main advantages of centralized AC systems is that they can satisfy the

6 cooling requirements for multiple buildings at the same time (Chow et al. 2004(b)). In

7 addition, they have the advantage of employing refrigeration equipment with large

8 capacity and high efficiency. Compared to split AC systems, the power required by

9 the equipment is lower (Chow et al. 2004(c); Shimoda et al. 2008; Soederman 2007;

10 Jordi et al. 2013). Moreover, for the usage of renewable energy sources, like

11 underground water or seawater, influenced by the type of cooling source, centralized

12 cooling systems are simpler and less expensive (Rezaie and Rosen 2012; Chow et al.

13 2004(a)). From an energy usage point of view and considering urban landscape,

14 centralized AC systems are effective and should be promoted.

15 However, in some respect, decentralized AC systems are more advantageous than

16 centralized systems. With a decentralized AC system, users have greater flexibility in

17 controlling the AC terminals according to their requirements. Under this type of

18 control method, the cooling energy supplied by the AC system would be reduced

19 effectively (Li and Jiang 2009). Moreover, there is no distribution system in

20 decentralized AC systems, which means that the total energy consumption would not

21 include the consumption of fans or pumps. Therefore, from the above analysis on

3
 1 system types, both centralized and decentralized AC systems have their own

2 advantages. From the comparison, it can be concluded that the centralized and

3 decentralized AC systems represent two entirely different AC concepts.

4 As Figure 1 shows, many studies (Li and Jiang 2009; Hu et al. 2004; Ren et al. 2003;

5 Long et al. 2003; Wu 2005; Chen et al. 2008; Ma et al. 2007; Li 2012; Building

6 Energy Research Center in Tsinghua University 2013; Sun 2006) have been

7 conducted to examine the energy consumption in residential buildings in different

8 districts of China. From the comparison, the annual energy consumption of

9 centralized AC systems is observed to be higher than that of decentralized systems in

10 general. The largest difference between the energy consumption of the two systems

11 could be greater than 10 times.

12
13 Figure 1. Comparison of annual energy consumption for decentralized and centralized AC
14 systems in residential buildings

15 Thus, the theoretical analyses of district cooling systems differ greatly from the actual

16 operating experience in many cases, and some important factors that lead to the large

17 differences in energy consumption between centralized and decentralized AC systems

18 have been ignored. In order to analyze the reason for the energy differences between

19 these AC systems, three actual engineering projects in residential buildings with

4
 1 different types of centralized AC systems were considered in this study. The

2 applicability of centralized AC systems in residential buildings is analyzed and

3 discussed, and the key elements influencing the energy consumption differences

4 between centralized and decentralized AC systems in the case studies are explored.

5 2 Methodology

6 The basic information of the three actual cases is graphically explained in Figure 2.

7 Centralized AC systems are applied in all the three cases; however, the levels of

8 centralization are relatively different.

9 AC systems can be considered as three heat transfer segments, namely (1) the heat

10 transfer process between AC terminals and indoor environment, (2) the chilled water

11 heat distribution process between refrigerating machines and AC terminals, and (3)

12 the cooling water heat distribution process between refrigerating machines and the

13 cooling side. In the segment of the heat transfer process between the AC terminals and

14 indoor environment, users in case 1 cannot adjust the AC terminals; however, in case

15 2 and case 3, users can turn on or turn off the AC terminals according to their

16 requirements. In the segment of the chilled water heat distribution process between

17 the refrigerating machines and AC terminals, in case 1 and case 2, all cooling energy

18 consumptions are centralized to the cold site and processed by a unified refrigeration

19 equipment with large capacity, while in case 3, household heat pump systems are

20 applied, and the heat pumps are distributed in each family. Because of different types

21 of refrigeration equipment, the types of chilled-water distribution systems are also

5
 1 different. In case 1 and case 2, the chilled water is supplied uniformly from

2 refrigeration plants to each AC terminal, while in case 3, there is no chilled-water

3 distribution process. In the segment of the cooling water heat distribution process

4 between refrigerating machines and the cooling side, all three cases use underground

5 water as the cooling source, and the cooling water is collected together for heat

6 dissipation. In case 3, however, the cooling water is circulated between the

7 underground cooling source and the heat pump in each household. Thus, the three

8 cases can be summarized as follows:

9 1. Case 1: centralized refrigerating machines and distribution system of chilled

10 water and cooling water; users cannot adjust the AC terminals;

11 2. Case 2: centralized refrigerating machines and distribution system of chilled

12 water and cooling water; users can adjust the AC terminals;

13 3. Case 3: decentralized refrigerating machines, centralized distribution system

14 of chilled water and cooling water; users can adjust the AC terminals.

15 It is difficult to get direct access to all three types of actual projects; therefore, we

16 seek assistance from other researchers. Case 1 (Li 2012) and case 3 (Sun 2006) have

17 been analyzed in previous studies, and the details of energy data are sufficient for our

18 analysis. Meanwhile, for case 2, we conducted field measurements in the community

19 to obtain the primary data for our study. The data collection procedure for each case

20 study is as follows:

21 1. Case 1: A metering system that records the cooling energy consumption by

6
 1 the community and the electricity consumption of the AC equipment (like

2 chillers, pumps, and fans) was installed in the AC system.

3 2. Case 2: An AC unit metering system that records the on/off state and the daily

4 run time of the fan coil unit (FCU) was installed in the community. The

5 cooling energy consumption was obtained from the property management

6 company in the residential community. Data (chilled-water flow rate from

7 each building, supply and return water temperatures for each building, chiller

8 flow rates, chiller supply and return water temperatures, and electrical power

9 to the chillers and pumps) were recorded every 10 min in the district cooling

10 system from July to September 2011. The energy consumption and system

11 efficiency were then calculated based on these measurements.

12 3. Case 3: Detailed measurements on the AC system in the community were

13 conducted, including the cooling consumption and the electricity consumption

14 of chiller and pumps. Moreover, the supply and return water temperatures of

15 the AC system were also recorded.

16 In the following discussions, the energy consumption of each segment and the reason

17 for the high-energy consumption of the centralized AC systems in residential

18 buildings are analyzed based on the measured data.

7
 Electricity
input

Heat Pump

Cooling water pump Chilled water pump

Compressor

Users

Heat Condenser Evaporator

exchanger
throttle

Cooling water Chilled water

distribution Heat pump distribution Terminals

Case 1 Centralized Centralized Centralized Un-adjustable

Case 2 Centralized Centralized Centralized Adjustable

Case 3 Centralized Decentralized × Adjustable

1

2 Figure 2. Research cases

4 3 Survey and measurement

5 2.1 Case 1: centralized chillers and distribution system; users cannot adjust

6 terminals

7 Case 1 is in a residential community in Jiangsu Province (Li 2012). There are 10

8 residential buildings in the community, and the total building area is 114,000 m2. The

9 community includes 1200 households, while the occupancy ratio is approximately 90%

10 at the time of measurement.

11 The system diagram of case 1 is shown in Figure 3. The AC terminals in this

12 community adopt ceiling panel radiation and displacement ventilation systems to meet

13 the requirements of temperature and relative humidity. Ground-coupled heat pump

14 units are chosen as cooling source. Variable speed pumps are employed on the

8
 1 cooling-water and chilled-water sides. The design cooling load of the ceiling panel

2 radiation system is 1757 kW, and two heat pumps are installed with unit number

3 control.

4 The typical totally centralized AC system is adopted in this residential community.

5 The chilled water collects the cooling consumption at the users’ side together and

6 brings it to the heat pumps to obtain the unified management. On the users’ side, the

7 cooling utility is charged by area. Each space in the buildings, including the corridors,

8 washrooms, and empty rooms, is controlled 24 h a day according to a predetermined

9 temperature and humidity standard, in order to keep the comfort requirement (24 °C

10 dry-bulb temperature) in every room all the time. Therefore, the service concept in

11 this community is centralized AC control. However, under this type of control mode,

12 the adjustability of the indoor environment by the users is considerably limited. For

13 example, in this community, users cannot open the windows to get fresh air, and the

14 AC terminals cannot be closed by the residents.

Electricity
input

Heat Pump

Cooling water pump Chilled water pump

Compressor

Users

Heat Condenser Evaporator

exchanger
throttle

Cooling water Chilled water

distribution Heat pump distribution Terminals

Centralized Centralized Centralized Un-adjustable

15

16 Figure 3. System diagram of case 1

9
 1 Under this type of service, the energy consumption of the AC system in this

2 community is shown in Figure 4. The split AC system is taken as the typical

3 representative of a decentralized AC system, and its consumption is also shown in

4 Figure 4 for comparison. The electricity consumption of the split AC system is

5 obtained from actual measurements of 780 households in Shanghai. The average

6 electricity consumption is 4.3 kWh/m2 (Long et al. 2003). Owing to the similarity of

7 the climate in Jiangsu and Shanghai, it is considered that the energy consumptions of

8 the split AC systems in the two districts are similar. From the comparison, the

9 electricity consumption of the AC system in this residential community is

10 approximately five times of that in the split AC system in the similar district.

11

Chillers Pumps FAU

25
2
Electricity consumption

20
（kWh/m2）

6.5
15

10
4.3
13.4
5

0
Case 1 Split AC system
12
13 Figure 4. Electricity consumption of AC system in case 1

14

15 Through the analysis of the measurements and survey results, it is found that the main

16 reason for the high AC consumption in this community is that under this type of

10
 1 system, namely, the non-adjustable feature for users, the AC operation mode is full

2 time and full space. Under this type of operation mode, the operating time and space

3 of the AC system under full load is considerably higher than that of the split AC

4 system. Under the centralized AC system in case 1, the product of the service area and

5 service time of the AC is 114,000 m2 × 24 h, which is equal to 2.740,000 m2·h. When

6 the split AC system is adopted, the product of the service area and service time is

7 approximately 684,000 m2·h, which is only 25% of that in case 1 (Building Energy

8 Research Center in Tsinghua University 2013). The efficiency of the heat pump

9 equipment itself is high, and the average COP in the cooling season is 4.4. However,

10 with the large requirement of cooling consumption in the terminal sides, the

11 electricity consumption of the heat pump itself is approximately two times of that in

12 the split AC systems.

13 Compared with split AC systems, the centralized AC system includes the electricity

14 consumption of water pumps and fans. According to the system operation records

15 from May 2009 to September 2009, the monthly electricity consumption of chillers,

16 water pumps, and fresh air units (FAU) is calculated as shown in Figure 5. The

17 proportion of electricity consumption for water pumps and fans in each month ranges

18 from 30% to 60%, which is approximately 0.5 to 1.4 times of the chiller consumption.

19 Therefore, the electricity consumption of the distribution system is a main

20 contributing part in the AC system and is a major factor in the high AC consumption

21 in this residential community.

11
 1

Chillers Pumps FAU

7

Electricity consumption
6
5
（kWh/m2） 4
3
2
1
0
2009.5 2009.6 2009.7 2009.8 2009.9
2

3 Figure 5. Composition of AC system's monthly energy consumption

5 Table 1. Composition of AC system electricity consumption in Case 1

Distribution Heat pump energy Distribution Cooling consumption

of cooling water consumption of chilled water
Case 1 2.5 kWh/m2 13.4 kWh/m2 4 kWh/m2 59 kWh/m2

7 The electricity consumption for each part of the centralized AC system in case 1 is

8 listed in Table 1. The efficiency of the entire AC system is equal to the users’ cooling

9 consumption divided by the sum of the electricity consumption of the heat pumps and

10 water pumps. Through calculation, we can obtain the system efficiency of this AC

11 system in case 1, which is 3.0. This value is higher than the typical efficiency of split

12 AC systems (2.5). However, in case 1, the operation mode is full time and full space,

13 and the AC terminals are non-adjustable. The actual cooling consumption of users in

14 this residential community is more than five times of that with the split AC systems in

15 the same district. Under high cooling requirement, the electricity consumption of the

12
 1 centralized AC system operation in case 1 is significantly higher than that of split

2 systems, and the distribution system consumes approximately 33% of the total

3 electricity consumption. With the two affecting elements, notwithstanding the high

4 efficiency of heat pump itself and the entire AC system, the electricity consumption of

5 the AC system in case 1 is approximately five times of the consumption when split

6 AC systems are adopted in the same district.

8 2.2 Case 2: centralized chillers and distribution system, users can adjust

9 terminals

10 Case 2 is in a residential community in Henan Province. The community covers an

11 area of 27,944 m2 with a habitable floor area of 41,200 m2. There are 12 buildings

12 with 294 households in the community, and each building has five floors. The

13 occupancy rate is 75%. Two screw water source heat pump units with cooling

14 capacities of 1463 kW and 542 kW, separately, are taken as the central AC equipment.

15 The pumps in the water system are constant-speed—the pumps include two

16 circulating pumps and two submersible pumps. The submersible pumps are working

17 under the water to pumps the water up. The list of main equipment is presented in

18 Table 2. The FCUs are taken as AC terminals with no control on the waterside. In this

19 community, users are charged for the cooling consumption according to the operation

20 situation of the FCU terminals, i.e., the cost of cooling consumption is based on the

21 measured operating hours of fan coils under high, middle, and low speeds. When the

13
 1 AC terminal is closed, there would be no cooling fees.

3 Table 2. Equipment list for case 2

Equipment Rated parameters

Rated cooling capacity: 1463.6 kW, rated cooling power:
Heat pump 1#
226.0 kW
Rated cooling capacity: 542.4 kW, rated cooling power:
Heat pump 2#
89.7 kW
Rated flow rate: 190 m³/h, pump head: 38 m, rated power:
Chilled water pumps
27 kW
Cooling water pumps 1#、2# rated power: 37 kW; 3#、4# rated power: 22 kW

5 The configuration of the AC system in case 2 is similar to that in case 1, as shown in

6 Figure 6. However, in case 2, the AC terminals are the FCUs, and users can choose

7 different speeds of the fan to adjust the indoor environment according to their

8 requirements. Moreover, the AC terminals can be closed to stop the AC supply in

9 some rooms. Therefore, the adjustability of this type of AC terminals is very similar to

10 that of the split AC system. As previously mentioned, in this community, the charging

11 method is based on the operating hours of the AC terminals, which encourages the

12 users to turn off the AC when not required. However, for the centralized refrigerating

13 machines and the constant-flow operation mode of the chilled-water system, the

14 electricity consumption of the chilled-water pumps takes a larger proportion of the

15 total energy consumption.

14
 Electricity
input

Heat Pump

Cooling water pump Chilled water pump

Compressor

Users

Heat Condenser Evaporator

exchanger
throttle

Cooling water Chilled water

distribution Heat pump distribution Terminals

Centralized Centralized Centralized Adjustable

1

2 Figure 6. System diagram for case 2

4 In case 2, the measured result of the users’ cooling consumption during the cooling

5 season is 7.5 kWh/m2. Under the air-conditioning system in case 2, the adjustability of

6 the AC terminal is similar to that of the split AC system; thus, it is assumed that the

7 cooling consumptions in the two types of AC systems are similar. Through simulation,

8 it can be calculated that if the totally centralized AC system (like the full-time and

9 full-space operation mode in case 1) is applied for the community in case 2, the

10 cooling consumption would increase to 54.0 kWh/m2, which is approximately seven

11 times of the actual existing cooling consumption. Therefore, when users can adjust the

12 AC terminals freely and a reasonable charging method is applied, the cooling

13 requirement of the users would be reduced greatly as compared to that in case 1.

14 It is supposed that when the split AC system is applied in this residential community,

15 the cooling consumption is the same as the actual cooling consumption in case 2.

16 Using the average COP of split AC systems as 2.5, then, we can calculate that when

15
 1 the split AC system is adopted, the electricity consumption of the AC system is 3.0

2 kWh/m2. This result is compared with the actual situation, as shown in Figure 7, and it

3 can be observed that the electricity consumption of the AC system in case 2 is more

4 than two times of that in the split AC system.

Chillers Pumps
8
Eletricity consumption

6 3.2
（kWh/m2）

4
3.0
2 4.4

0
Case 2 Split AC system
5

6 Figure 7. Electricity consumption of AC system in case 2

8 Figure 8 shows the service time proportion of the FCU terminals, which is equal to

9 the total operating time in each household divided by the total number of FCUs in the

10 community and the operation hours of the heat pump. The operation data of the

11 household FCUs are obtained from the household cooling consumption meter system.

12 The results show that nearly 80 households did not open the FCU terminals during the

13 observation period, and more than 1/3 of the users indicated the use of FCU terminal

14 service time at less than 10%. The calculation showed that the average proportion of

15 the FCU terminal service time of this residential community during the entire cooling

16 season is only 7%.

16
 60%

50%
Open Rate of FCU

40%

30%

20%

10%

0%
1 21 41 61 81 101 121 141 161 181 201 221 241 261 281
Users
1

2 Figure 8. Proportion of FCU terminal service time from July 1 to 20, 2011

4 Under this AC terminal operation mode, the contemporary usage ratio of the AC is

5 low, which resulted in a low load ratio of the entire residential community. With the

6 small load ratio, the efficiency of the entire AC system is extremely low. The system

7 efficiency in case 2, which is equal to the total cooling consumption divided by the

8 total electricity consumption, is approximately 1 (7.5/7.6). The main reasons for this

9 low system efficiency are as follows:

10

11 1. On one hand, as mentioned above, under the part-time, part-space AC operation

12 mode in this residential community, the load ratio of the entire system is

13 relatively low. Only few users have high cooling load requirements, while

14 majority of users have low or no cooling load requirements. Therefore, most of

15 the time, only few fan coils would operate in the AC system. However, the water

17
 1 system has a constant flow rate— the water system would work under a

2 situation with large water flow rate and small differences between the supply

3 and return water temperature. As Figure 9 shows, during the entire cooling

4 season, the chilled-water temperature difference between the supply and return

5 side of each building is only approximately 1 °C. Therefore, the electricity

6 consumption of the water pumps occupies a larger proportion. The electricity

7 consumption of the water pumps can account for 42% of the total electricity

8 consumption.

9 2. On the other hand, because of the low cooling load requirement in the community,

10 the heat pumps work under low load ratio for a long time. The COP of the heat

11 pumps is considerably lower than the rated value of 6.4. During the entire

12 cooling season, the average COP of the heat pumps is only 1.7, and the system

13 efficiency of the entire AC system is only 1. However, even if the COP of the

14 heat pump can reach the rated value of 6.4, the system efficiency can only

15 increase to 1.72. The improvement of system efficiency is limited by the

16 electricity consumption of water pumps. From these analyses, it can be

17 concluded that the electricity consumption of the water pumps is the main factor

18 that limits the increase of the system efficiency of the entire AC system.

18
 Distribution of T differences beween supply and return side
Temperature differences(℃) 1.8
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
1# 2# 3# 4# 5# 6# 7# 8# 9# 10# 11# 12#
1
2 Figure 9. Chilled-water temperature differences between supply and return side of each building

4 In conclusion, through the analysis of case 2, when users have the ability to control

5 the AC terminals, the cooling consumption from the users’ side would reduce greatly

6 compared with the centralized AC system in case 1. Influenced by the system type,

7 the electricity consumption of the AC system in case 2 includes the electricity

8 consumption of the water pumps. With the constant flow rate operating mode of the

9 water system, the electricity consumption of the water pumps becomes the main

10 contributor to the total AC consumption and is also the main reason for the low

11 system efficiency.

12 2.3 Case 3: decentralized AC system, centralized distribution system; users can

13 adjust the AC terminals

14 Case 3 is in a residential community in Beijing (Sun 2006). There are 3 residential

15 buildings in the community, and the total cooling area is 70,000 m2. The community

16 includes 368 households. The distributed water loop heat pump system is adopted as

19
 1 the AC system，and the design cooling load is 64 W/m2. As shown in Figure 10, the

2 AC system includes a centralized underground water circulating system and a

3 decentralized heat pump system. The underground water is lifted by a deep-well pump.

4 After heat exchange, cooling water is supplied to the heat pump in each household

5 through the loop pipe network. The heat pump is distributed in each household. In

6 order to satisfy the cooling requirement, the circulating pumps operate 24 h a day

7 continuously under a constant flow rate. The cooling water pump and water collection

8 pump are variable-speed.

Building 1 …… Building n Case 3

Heat pump Heat pump Heat pump

Heat pump Heat pump Heat pump

Decentralized
Users’ side

Heat pump Heat pump Heat pump

Distribution of
cooling water
Cooling water pump Centralized
Heat
exchanger
Water collection
pump

9
10 Figure 10. System diagram of case 3

11

12 An all-air system is taken as the AC terminal in this community. Air ducts are

20
 1 installed in each room, and there is no control valve in the air inlet. Therefore, when

2 there is a cooling requirement in a room, all rooms in the household would be cooled.

3 In contrast with the two cases above, the heat pump is set in each household in case 3,

4 and the heat pump can be turned on or off according to the requirement. Thus, the

5 operation mode is part time, full space. In case 3, the AC system still has a centralized

6 feature, i.e., the cooling water circulates uniformly, and the AC terminals in each

7 household are controlled uniformly. Users will pay for the electricity consumption of

8 heat pumps and a portion of other consumptions is charge based on the area of the

9 household.

10 Researchers have obtained the average electricity consumption of the split AC system

11 as applied in China. The average value is 3.1 kWh/m2 in residential buildings in

12 Beijing (Li 2006). The electricity consumption of the AC system in the residential

13 community in case 3 is approximately three times of that in the split AC system, as

14 shown in Figure 11. Considering the average COP of the split AC system as 2.5, the

15 average cooling consumption of the split AC system in residential buildings in Beijing

16 is approximately 7.8 kWh/m2. The measured cooling consumption in case 3 is 13.2

17 kWh/m2. Therefore, under this type of system, the cooling consumption is higher than

18 that of the split AC system. The main reason for this is that under the split AC system,

19 the usage mode is part time and part space, while in case 3, which is influenced by the

20 AC terminals, the usage mode is part time and full space.

21
 10

Electricity consumption
1.2 Chillers
8 Cooling circulating pumps

（kWh/m2）
3.6 Water collection pumps
6

4
3.1
2 4.6

0
Case 3 Split AC system
1
2 Figure 11. Electricity consumption of AC system in case 3

4 During the entire cooling season, the average COP of the heat pumps is 2.9, which is

5 higher than the split AC system. However, in case 3, all the water pumps are turned on

6 for the entire day, irrespective of the number of heat pumps under operation. Only the

7 circulating pumps have an electricity consumption of 3.7 kWh/m2. Table 3 shows the

8 average result of the temperature measurement during the entire cooling season. From

9 the measured result of cooling water temperature differences, it can be observed that

10 the water system is under a situation with large water flow rate and small differences

11 between the supply and return water temperatures as shown in Table 3. Under this

12 operating situation, the consumption by the water pumps becomes the main

13 composition of the total electricity consumption, which can be observed in Figure 11.

14 In this cooling water centralized AC system, the electricity consumption of water

15 pumps is approximately half of the total consumption. Therefore, the electricity

16 consumption of AC system is higher than that of general split AC systems.

17

22
 1 Table 3. Cooling water temperature differences between supply and return side

Tsupply Treturn T differences

Design value 18 32 14
2003 25.8 26.5 0.7
Measured data of each year 2004 27.9 28.5 0.6
2005 27.0 27.8 0.8

3 In conclusion, case 3 shows that under a cooling supply mode of part time and full

4 space, the cooling requirement of users would be less than that of the total centralized

5 AC system (case 1); however, it is higher than the cooling consumption under the part

6 time and part space service mode. In case 3, the system adopts a decentralized heat

7 pump, which ensures that the heat pumps can work under a high load ratio; therefore,

8 the operating performance of the heat pumps is improved. However, the cooling water

9 system is centralized, and the electricity consumption of the cooling water distributed

10 system becomes the main part of the total consumption. The electricity consumption

11 of the cooling water pumps is approximately 51% of the total electricity consumption

12 of the AC system.

13

14 3 Analysis and discussion

15 The three case studies discussed above all adopt centralized AC systems; however,

16 they are different in the users’ ability to control the AC terminals, type of refrigeration

17 equipment, and form of distribution system. Taking the three cases together and

18 making a comprehensive comparison will enhance the understanding of the

23
 1 applicability of centralized AC systems in residential buildings. For the differences in

2 thermal load, service area, room temperature, etc., it is not appropriate to compare the

3 energy consumption of the three cases directly. Therefore, in this study, we focus on

4 the characteristics of each case, and only the ratios of each energy consumption part

5 are compared.

7 Table 4. Comparative analysis of three cases

Distribution of Heat pumps Distribution of Cooling

cooling water chilled water consumption
Case 1 Centralized Centralized Centralized Un-adjustable
2.5 kWh/m2 13.4 kWh/m2 4 kWh/m2 59 kWh/m2
Case 2 Centralized Centralized Centralized Adjustable
1.7 kWh/m2 4.4 kWh/m2 1.5 kWh/m2 7.5 kWh/m2
Case 3 Centralized Decentralized - Adjustable
4.8 kWh/m2 4.6 kWh/m2 13.2 kWh/m2

9 Table 5. Efficiency comparative analysis

Transport
Transport Electricity
coefficient of COP of heat System
coefficient of consumption
cooling pumps efficiency
chilled water kWh/m2
water
Case 1 29.0 4.4 14.8 3.0 21.9
Case 2 7.0 1.7 5.0 1.0 7.6
Case 3 3.7 2.9 - 1.4 9.4

10

11 The composition of the electrical and cooling consumption in the three cases are listed

12 in Table 4, and Table 5 lists the energy efficiency of each segment. The calculation

13 method of energy efficiency is based on the requirement in the Chinese standard for

24
 1 “economic operation of AC control system.” The COP is equal to the cooling

2 consumption divided by the electricity consumption of the heat pumps; the transport

3 coefficient of chilled water (TCchw) is equal to the cooling consumption divided by the

4 electricity consumption of chilled-water pumps, while the transport coefficient of

5 cooling water (TCcdp) is equal to the cooling consumption divided by the electricity

6 consumption of cooling water pumps; the system efficiency is equal to the cooling

7 consumption divided by the sum of the electricity consumption of heat pumps and

8 water pumps. The electricity consumption is the sum of the consumptions of all

9 components (cooling water pumps, chilled-water pumps, and heat pumps). Through

10 the comparison and analysis, the main conclusions from the three study cases are as

11 follows:

12 1. If there is a condition for independent adjustment, users would adjust the AC

13 terminals according to the operation of split AC system. The cooling load would

14 express features of desynchrony and low load ratio.

15 In cold and extremely cold areas, the cooling and heating loads in residential

16 buildings have different characteristics. During the heating season, the heating load is

17 mainly influenced by the outdoor climate; thus, the heating loads of different users are

18 synchronous. However, during the cooling season, the internal heat gain becomes the

19 main influencing element of the cooling load, resulting in different characteristics of

20 heating and cooling loads.

21 Apartment is the main building type in Chinese residential buildings. The main

25
 1 features of the operation mode in apartments are: 1) the occupancy situation differs

2 among different households; 2) the number of people in the household changes

3 substantially; 3) the requirements of indoor environments have large discrepancies.

4 With these features, if users could control the AC terminals freely, they would operate

5 the AC part time and part space, and the operation mode results in the load feature of

6 desynchrony and low load ratio in residential buildings.

7 Through the comparison, we can find that in totally centralized AC systems, the full

8 time and full space indoor environment control service (case 1) results in the highest

9 electricity consumption. The energy consumption in case 1 is approximately three

10 times of that in the situation where the AC terminals can be controlled freely (case 2).

11 The service mode is one of the most important factors affecting the differences in

12 energy consumption. In case 1, the cooling consumption is 59 kWh/m2, which is five

13 times higher than that in case 2 and case 3. In case 1, no matter what the cooling

14 requirement is, the indoor environment is controlled according to the predefined

15 standard of comfort, i.e., the full time and full space indoor environment control mode.

16 In case 3, the household centralized AC system is adopted, and the service mode is

17 part time and full space. In this case, the cooling consumption is 13.2 kWh/m2, which

18 is lower than that in case 1. In case 2, the control of AC terminals is similar to that in

19 split AC systems, and users can control the AC terminals according to their

20 requirements. Under this type of control mode, influenced by the lifestyle, most

21 residents in China would select the part time and part space operating mode.

26
 1 Therefore, the cooling consumption in case 2 is lower compared to that in case 1 and

2 case 3. For example, the bedroom in a household will only be occupied during 30% of

3 the day, and the residents would like to close the air conditioner after they fall asleep.

4 Thus, the time when the air conditioner is opened in the bedroom is only

5 approximately 20% of the day. However, the same bedroom in case 1 would be air

6 conditioned all day long; therefore, the cooling consumption would be considerably

7 higher than the actual requirement. Under this type of service mode, the factor of

8 service time and space is almost five times of that in split AC systems.

10 2. In residential buildings, influenced by the requirement of user terminals, the

11 distribution energy consumption becomes the main component of the system

12 consumption in centralized AC systems.

13 In case 2 and case 3, the control of AC terminals possesses the feature of that in split

14 AC systems; however, the refrigerating machines and distribution systems are

15 conventional types in centralized AC systems. In case 2 and case 3, the energy

16 consumption of the AC system is approximately 8 kWh/m2, which is approximately

17 1.5 to 3 times of those in decentralized AC systems.

18 This difference in energy consumption is mainly caused by the distribution energy

19 consumption in case 2 and case 3. From Table 5, it can be observed that the transport

20 coefficient of chilled water and cooling water cannot exceed 10. In case 3, the

21 transport coefficient of chilled water is only 3.7. According to the stipulation in the

27
 1 Chinese standard for “economic operation of air-conditioning control system,” when

2 the annual working condition is evaluated, the lower limit value for TCchw and TCcdp is

3 30 and 25, respectively. Therefore, in case 2 and case 3, the transport coefficients of

4 the pumps are extremely low.

5 As mentioned above, in case 2 and case 3, the cooling consumptions do not differ

6 considerably with those in decentralized AC systems. However, the circulating pumps

7 in the distribution systems continuously operate for 24 h; thus, most of the time, the

8 efficiency of distributing the cooling consumption is low. For example, in case 2,

9 during the cooling season, the circulating pump electricity consumption alone is 3.2

10 kWh/m2, and this value is equal to the average electricity consumption of the split AC

11 system in summer in that area. Case 3 can explain the phenomenon further. In case 3,

12 only the cooling water is circulating in the system. With the ability to control the AC

13 terminals, residents will take the control mode of part time and part space; thus, the

14 operating time of the heat pumps will be reduced considerably. From the results in

15 case 3, the electricity consumption of heat pumps and terminal equipment is only 48%,

16 and the electricity consumption of cooling water pumps is more than half of the total

17 consumption.

18 If electric on/off valves are installed in the users’ side, and the water pumps use

19 frequency conversion control in case 2 and case 3, then the water temperature

20 differences of the chilled-water system can be increased, and this can reduce the

21 distribution system’s energy consumption in the centralized system to a certain extent.

28
 1 However, with desynchrony and low load ratio feature of the cooling load in

2 residential buildings, the energy consumption of the distribution system will still be a

3 major component of the total consumption. Taking case 2 as an example, Figure 12 is

4 the statistical result of each household’s AC terminal operating time frequency on July

5 4. Majority of users’ operation ratio of FCU terminals is within 10%, and only a few

6 users’ operation ratio reaches 60%. Under this situation, the most ideal control mode

7 of the water system is that many circulating pumps are installed using parallel

8 operation control, and the operation number and frequency of the water pumps are

9 adjusted according to the water temperature differences. However, in actual

10 engineering cases, in general, the number of pumps would be two to three. Therefore,

11 when the AC terminal operation ratio is low, the water pump operating point would

12 have a serious deviation that will result in high consumption and low efficiency of the

13 distribution system. Meanwhile, the chillers also limit the reduction of the energy

14 consumption of the distribution systems. In centralized AC systems, there are two to

15 three chillers at most, and each chiller has a minimum flow restriction. Therefore, the

16 flow rate of the distribution system cannot be reduced considerably under low load

17 ratio. Consequently, the system will inevitably operate under a situation with large

18 water flow rate and small differences between the supply and return water

19 temperatures.

20 If a variable-speed pump control is applied in case 2, the electricity consumption of

21 the water pumps can be reduced. However, it is relatively difficult to handle a

29
 1 situation where the cooling consumption is extremely low and unevenly distributed

2 among different households (as Figure 8 shows). Other pumping systems, like

3 primary-secondary, as well as primary-tertiary systems may mitigate the problem,

4 which requires the introduction of simulation methodology and further analysis.

5 Therefore, from the above analysis, the main reason for the low system efficiency is

6 that the cooling load in residential buildings has the feature of desynchrony and low

7 load ratio, and this feature considerably differs from that in public buildings. During

8 the design phase, the engineering load calculation method usually focuses on the peak

9 load for equipment sizing, while load diversity is rarely considered. In addition, even

10 when the hourly cooling load is calculated, the internal load and schedule of each

11 household are the same; thus, the disparity due to desynchrony and low load ratio

12 would not be considered.

13

60%
Operation time frequency

50%
40%
30%
20%
10%
0%
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Open ratio
14

15 Figure 12. Statistical result of operating time frequency of FUC terminal on July 4 in case 2

16

17 3. In residential buildings, the sections where the centralized feature meets the

30
 1 decentralized feature frequently present high energy consumption characteristics.

2 From the above analysis, the cooling requirement in residential buildings has the

3 feature of part time and part space. This feature is in conflict with the inherent

4 disadvantage of adjustability in centralized AC systems. In actual applications, when

5 adjustability is limited in AC terminals or distribution systems, the system would

6 frequently present a high energy consumption.

7 In case 1, the AC terminals cannot be adjusted, and this is in conflict with the

8 decentralized feature of the cooling load on the users’ side. This contradiction leads to

9 an increase of cooling consumption, which results in the high operation consumption

10 in case 1.

11 In case 2, the AC terminals are adjustable; however, the water system has a constant

12 flow rate, which is in conflict with the users’ decentralized cooling load. Particularly,

13 in case 2, the heat distribution process of chilled water between the refrigerating

14 machines and AC terminals becomes the contradiction boundary. This contradiction

15 makes the distribution and consumption of chilled-water pumps the main reason for

16 high energy consumption. From Table 5, the transport coefficient of chilled water in

17 case 2 is only 5, which is significantly below the standard limit value of 30. The

18 non-adjustability of the chilled-water system influences the adjustability and

19 operating condition of chillers and cooling water systems, resulting in low system

20 efficiency.

21 In case 3, the household heat pump system is adopted; therefore, the heat pumps and

31
 1 the indoor heat transfer can be taken in entirety and possess the decentralized feature.

2 However, the cooling water system has a constant flow rate. Consequently, this

3 contradiction makes the electricity consumption of the cooling water pumps as the

4 main component of the total consumption. Comparing case 2 and case 3, where both

5 heat pumps are adopted, the electricity of the cooling water pumps in case 2 is only 20%

6 of the total consumption, while the consumption of the cooling water system in case 3

7 is 51%.

8 Therefore, through the comparison of the three cases, when a centralized AC system

9 is applied in residential buildings, it is important to increase the adjustability of each

10 segment to match the decentralized load characteristics. If any segment lacks

11 adjustability, it would become a contradiction boundary, and becomes a point where

12 high consumption or low efficiency easily occurs.

13 Therefore, from the comparison, it can be concluded that the AC usage mode,

14 distribution system, and adjustability greatly influence the operation of the centralized

15 AC system. Particularly in residential buildings where the cooling load has the feature

16 of desynchrony and low load ratio, it is difficult to predict during the design phase the

17 correct sizing, control method, and pumping method that fit all types of situations.

18 This would be an extremely difficult task and would require a detailed analysis.

19 However, it should be noted that other elements, like architectural characteristics and

20 climate would also influence the performance of the AC system. More research must

21 be conducted to determine the degree of influence of AC system types on the energy

32
 1 consumption and system efficiency in the near future.

3 4 Conclusion

4 The data from the three cases are analyzed and discussed in this study and the

5 following conclusions are presented:

6 1. When centralized AC systems are applied in residential buildings, three actual

7 engineering cases indicate that the adjustability in each segment (users’

8 terminal, refrigeration equipment, and distribution system) would greatly

9 influence the energy consumption and system efficiency.

10 2. The inherent load feature of desynchrony and low load ratio would influence

11 the application of centralized AC systems in residential buildings.

12 3. The design scheme should improve the system operation efficiency on the

13 premise of realizing the adjustability of the system.

14

15 Acknowledgements

16 This work was supported by the Key Projects in the National Science & Technology

17 Pillar Program during the 13th Five-year Plan Period (No. 2016YFC0700102) and

18 Jiangsu Natural Science Foundation (BK20160683).

19

20 References

21 Aste N, Adhikari RS, Manfren M (2013). Cost optimal analysis of heat pump

33
 1 technology adoption in residential reference buildings. Renewable energy, 60:

2 615-624

3 Building Energy Research Center in Tsinghua University (2013). 2013 annual report

4 on China building energy efficiency. Beijing, China building industrial publishing

5 (in Chinese)

6 Chen SQ, Li NP, Guan J, Xie YQ, Sun FM, et al (2008). A statistical method to

7 investigate national energy consumption in the residential building sector of China.

8 Energy and Buildings, 40(4): 654-665.

9 Chow TT, Au WH, Yau R (2004a). Applying district-cooling technology in Hong

10 Kong. Applied Energy, 79(3): 275-289

11 Chow TT, Chan ALS, Song CL (2004b). Building-mix optimization in district-cooling

12 system implementation. Applied energy, 77: 1-3

13 Chow TT, Fong KF, Chan ALS (2004c). Energy modeling of district cooling system

14 for new urban development. Energy and Buildings, 36(11): 1153-1162

15 Hu PF, Jiang ZN, Leng YH, Xiang CW (2004). Investigation of thermal environment

16 and energy consumption for Hubei residences. Heating ventilation & Air

17 Conditioning, 34(6): 21-22, 71 (in Chinese)

18 Jordi O, Joan CB, Alberto C (2013). Operational optimisation of a complex

19 trigeneration system connected to a district heating and cooling network. Applied

20 thermal engineering, 50(2): 1536-1542

21 Li Z (2012). Investigation and analysis of the relationship between residential energy

34
 1 use behavior and energy consumption in China. PhD Thesis, School of architecture in

2 Tsinghua University, Beijing (in Chinese)

3 Li ZJ, Jiang Y (2009). Analysis on cooling energy consumption of residential

4 buildings in China’s urban areas. Heating ventilation & Air Conditioning, 39(5):

5 82-88 (in Chinese)

6 Long WD, Zhong T, Zhang BH (2003). Situation and trends of residential building

7 environment services in Shanghai. Proceedings of the 2003(4th International

8 Symposium on Heating, Ventilating and Air Conditioning. Beijing: Tsinghua

9 University Press: 493-498

10 Ma BQ, Yan ZF, Gui ZG, Tan W, He JC (2007). Investigation and analysis on summer

11 energy consumption structure of energy efficient residential buildings in Xi’an.

12 Building Science, 23(8): 53-56, 60(in Chinese)

13 Ren J, Meng QL, Liu Y, Yang SR (2003). Analysis and research about AC energy

14 consumption of residential buildings in Guanzhou. Walling Material Innovation and

15 Energy Conservation of Buildings, 4: 34-37 (in Chinese)

16 Rezaie B, Rosen MA (2012). District heating and cooling: Review of technology and

17 potential enhancements. Applied energy, 93(SI): 2-10

18 Shimoda Y, Nagota T, Isayama N, Mizuno M (2008). Verification of energy efficiency

19 of district heating and cooling system by simulation considering design and operation

20 parameters. Building and Environment, 43(3): 569-577

21 Soederman J (2007). Optimisation of structure and operation of district cooling

35
 1 networks in urban regions. Applied thermal engineering, 27(16):2665-2676

2 Sun F (2006). Research of ground water-source heat pump air conditioning system.

3 PhD Thesis, Beijing University of Technology, Beijing (in Chinese)

4 Wu Q (2005). Study on the energy-consumption problem and energy-efficiency

5 technologies of residential buildings in Hangzhou. PhD Thesis, College of Civil

6 Engineering and Architecture in Zhejiang University, Zhejiang (in Chinese)

7 Zhang WJ, Huai LL, Liu JL (2009). Comparison analysis of energy consumption

8 characteristics between decentralized and centralized AC systems in residential

9 buildings by commissioning. Proceedings of the fifth international workshop on

10 energy and environment of residential buildings and third international conference on

11 built environment and public health, vol I and II: 1768-1775

12

36

View publication stats