- Email: [email protected]
A method to evaluate building energy consumption based on energy use index of different functional sectors
Journal Pre-proof A Method to Evaluate Building Energy Consumption Based on Energy Use Index of Different Functional Sectors Zhuling Zheng, Jialin Wu, Zhiwei Lian, Huibo Zhang
PII:
S2210-6707(19)31650-6
DOI:
https://doi.org/10.1016/j.scs.2019.101893
Reference:
SCS 101893
To appear in:
Sustainable Cities and Society
Received Date:
10 June 2019
Revised Date:
13 October 2019
Accepted Date:
13 October 2019
Please cite this article as: Zheng Z, Wu J, Lian Z, Zhang H, A Method to Evaluate Building Energy Consumption Based on Energy Use Index of Different Functional Sectors, Sustainable Cities and Society (2019), doi: https://doi.org/10.1016/j.scs.2019.101893
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2019 Published by Elsevier.
A Method to Evaluate Building Energy Consumption Based on Energy Use Index of Different Functional Sectors Zhuling Zheng1,2, Jialin Wu1 , Zhiwei Lian1,*, Huibo Zhang1 1
School of Design, Shanghai Jiao Tong University, Shanghai, China 200240
[email protected] [email protected] [email protected] *Corresponding email: [email protected] 2
ro of
Shanghai Jianke Building Energy Service Co., Ltd, Shanghai, China 200032
Highlights
The energy consumption index of multifunctional buildings calculated by the regression model was consistent with that of pure functional buildings. Therefore, the results of sub-item energy consumption index can also be used to evaluate the energy consumption level in pure functional buildings.
The deviation of sub-item energy consumption index was small between the multifunctional buildings and pure functional buildings.
The sub-item energy intensity of each functional zone in Shanghai was obtained, providing reference for fast building energy consumption prediction in Shanghai.
lP
re
-p
Abstract
Jo
ur
na
Appropriate management of energy consumption provides great support for the sustainable development of buildings and even the whole cities. Nowadays, comprehensive energy consumption monitoring index is widely used in many practical projects to make horizontal comparisons among buildings. However, problems may occur when this method is applied in comprehensive buildings. This paper proposed a method to evaluate the overall energy consumption of buildings based on the energy index obtained from different functional sectors. A linear regression model was developed to predict energy consumption index in multifunctional areas. The model was based on online monitoring data from 30 single functional buildings and 20 multifunctional buildings. The sub-item energy consumption intensity difference between pure hotel buildings and hotel sector in comprehensive buildings was within 4.2%. The deviation of sub-item energy consumption intensity was less than 12% between pure shopping malls and shopping mall sector in comprehensive buildings. As for the pure office buildings and office sector in comprehensive buildings, except for heat pumps and chilled water pumps, the deviations of energy intensity of other sub-items ranged from 3.9% to 16.7%. The results from on-site experiments show that the sub-item energy consumption index obtained from multifunctional buildings had lower error than that of single functional buildings. Therefore, the sub-item energy consumption index can also be used to evaluate energy consumption in pure functional buildings. The model proposed in this paper provides a good reference for the prediction of energy consumption index in large public buildings.
Keywords Energy consumption; Linear regression; Single functional buildings; Multifunctional buildings
1. Introduction
ro of
China has now become the second largest country in building energy consumption (Eom, Kyle, Clark, Patel & Kim, 2012). The energy consumption in public buildings accounted for about 26.4%, and the energy consumption per unit area increased from 16.5kgce/m2 in 2001 to 21.9kgce/m2 in 2012 at an average rate of 7% annually (IPCC Fifth Assessment Report, 2014). According to the statistics about the electricity consumption intensity of public buildings in China in recent years (China Statistical Yearbook, 2015), the growth of public building area slowed down, while electricity consumption intensity increased exponentially. Urban construction greatly created an environmental burden, triggering the acceleration of climate change (Sandanayake, Zhang & Setunge, 2018). In the construction process, it is much limited that new materials, equipment, systems and other technologies could be adopted to reduce energy consumption. Actually, appropriate management of energy consumption may also provide great support for sustainable development of buildings and even the whole cities. Therefore, intelligent energy management recently became the main source of energy conservation (Killy & Fokaides, 2015).
Jo
ur
na
lP
re
-p
However, with a large amount of online monitoring data obtained from energy consumption monitoring platforms, it is necessary to provide online energy-saving diagnosis techniques. Kontokosta and Tull (2017) used several algorithms, including linear fit, random forest, and support vector machine, to build a data-driven model based on the energy consumption data from 23,000 buildings and predicted the energy consumption of 1.1 million buildings in New York. Brownsword, Fleming, Powell and Pearsall (2005) established a fitting algorithm for the energy demand of different consumption types in Leicester according to survey data, and then the algorithm was used to predict short-term energy demand. Zhu et al. (2019) presented a systematic way of qualifying daily building load patterns and identifying abnormal energy consumption. These were real-time energy-saving diagnosis based on the energy consumption prediction. However, the algorithm itself was much complicated and the accuracy of the algorithm strongly depended on the model and parameter setting, which required enough historical data or a large number of building samples. Wang (2010) proposed three basic energy-saving diagnostic methods, including efficiency indicator method, and provided specific diagnosis cases. Zhang (2010) proposed a seasonal gray theory prediction algorithm by improving the smoothness and residual model, which was applied in energy conservation monitoring platforms of public buildings in Shanxi, China. Although these algorithms were simplified and applied into practice, only certain types of buildings were included and the monitoring platforms still cannot serve for most common types of buildings. Some scholars have studied the energy consumption quota in various types of buildings. In some research (Lee, 2010; Kinney, Piette & Berkeley, 2002; Thomas, Alan & Roberto, 2004), the evaluation methods of the total energy consumption in several buildings were given. EUI (Energy Use intensity) index or normalized EUI index was adopted to measure the overall energy use of the building. Chung, Hui and Lam (2006) introduced EUI index and found its influencing factors based on multiple regression analysis, and a benchmark electricity meter for EUI was obtained. In terms of Cal-Arch, one of the energy evaluation tools, the distribution of energy consumption in various buildings was explored from the CEUS database in the United States and Canada, and the
EUI index of each building was compared according to the distribution histogram to evaluate the building energy (Matson & Piette, 2005). Zhang (2009) attempted to find an energy consumption quota where 60% of building energy consumption was guaranteed to be less than this value, according to energy consumption survey and information collection. Hu (2018) obtained a reasonable quota level by regression analysis and energy consumption quotas, which ensured that energy consumption in most buildings is lower than the control level, providing a good basis for energy conservation supervision. However, several problems existed in this quota level. For example, overall energy consumption could meet the standard while the local part may exceed the standard. If the energy consumption among buildings cannot be objectively evaluated, energy-saving diagnosis and renovation in large-scale public buildings will be affected.
-p
ro of
Li (2008) proposed crossing method on sub-metering system platforms. Local diagnosis was first carried out and energy-saving potential was then comprehensively explored in public buildings. Zhao, Zhu and Wu (2009) analyzed the energy consumption of public buildings in detail based on energy-saving behavior, thermal environment and air-conditioning operation management, and methods for energy consumption diagnosis were also given. These strategies were used to save energy mainly from on-site operations of the equipment. Therefore, this paper proposes a regression model based on massive sub-item energy consumption data. The model also provides new thoughts for online energy-saving diagnosis.
re
2. Methods
Jo
ur
na
lP
2.1. Distribution of building energy consumption In order to find out the distribution of energy consumption in different functional buildings, this paper collected enough energy consumption data from 130 public buildings in Shanghai. The energy consumption intensity of each building is shown in Fig. 1.
Fig. 1. Total energy consumption intensity of 130 large public buildings.
According to the statistics, the total energy consumption per unit area ranged from 50.5kWh/(m2 · a) to 430.2kWh/(m2·a), and the difference was obvious. The total energy
ro of
consumption of shopping malls and supermarkets was the highest, with the average value of 271.6kWh/(m2·a) and 281.7kWh/(m2·a) respectively, and that of commercial and hotel buildings was 147.9kWh/(m2·a) and 142.7kWh/(m2·a) respectively. Energy consumption in office buildings was relatively low, with the average value of 107kWh/(m2·a). However, comprehensive commercial buildings occupy a large part of cities. Generally, shopping mall sectors are arranged at the lower floor of comprehensive buildings and the upper part is office sectors. And the disparity of energy consumption per unit area is large between shopping mall and office sectors. Therefore, the proportion of various functional areas is different, resulting in the difference of energy intensity among types of multifunctional buildings. Reasonable evaluation method about building energy consumption should be introduced. It can be found that even if the buildings include the same functional sectors, the area of these sectors will greatly affect the overall building energy consumption. Thus, this paper intends to establish an evaluation model based on the functional area.
re
-p
2.2. Regression model Some scholars have used multiple linear regression methods to study the sensitivity of design parameters, such as envelope, HVAC system and building materials, towards building energy consumption (Zhan, Li & Hu, 2012; Wang & Pan, 2014; Lam, Wan & Yang, 2008; Lam, Wan & Liu, 2010). Since this paper considers the impact of functional area on the overall energy consumption, each functional sector can be regarded as the “weight” affecting energy consumption. Therefore, linear regression model is also adopted in this paper for data fitting.
lP
It is assumed that yi be total energy consumption, where i=1, 2...n (n is the number of samples). Since the area of functional sectors greatly affects the overall energy consumption, it can be considered as an independent variable in the regression model. Si,j means the area of different functional sectors, where 𝑗 = 1, … , m (m is the type of functional sectors).
ur
na
Generally, functional sectors can be divided into the following seven categories: office, accommodation, shopping, sports, education, medical treatment and others. Four types of most common functional sectors are involved in this paper: office, hotel, shopping mall, garage, which are also covered by the selected building samples. Therefore, m is taken as 4.
Jo
In order to find the relationship between functional area and total energy consumption and obtain the basic parameters, general linear regression model is proposed as equation (1). 𝑦𝑖 = 𝛼0 + 𝛼1 𝑆𝑖,1 + 𝛼2 𝑆𝑖,2 + 𝛼3 𝑆𝑖,3 + 𝛼4 𝑆𝑖,4 + 𝜀𝑖 (1) Where, yi——sub-item energy consumption, kWh/d; α1 ——sub-item energy consumption intensity of hotel sector, kWh/(m2·d); α2——sub-item energy consumption intensity of office sector, kWh/(m2·d); α3——sub-item energy consumption intensity of shopping mall sector, kWh/(m2·d); α4——sub-item energy consumption intensity of garage sector, kWh/(m2·d); Si,1——area of hotel sector, m2; Si,2——area of office sector, m2; Si,3——area of shopping mall sector, m2; Si,4——area of garage sector, m2;
εi——error, kWh/d. Energy consumption intensity is an important index to evaluate energy consumption in certain functional sector. Sub-item energy consumption intensity in certain functional sector is defined as equation (2):
j=
W
j
(2)
Sj
Where, αj——sub-item energy consumption intensity, kWh/m2; Wj——sub-item energy consumption in certain functional sector, kWh; Sj——area of certain functional sector, m2.
ro of
Since the total energy consumption involves each functional sector, it is assumed that 𝛼0 = 0. Equation (1) can be further simplified to equation (3): 𝑦𝑖 = 𝛼1 𝑆𝑖,1 + 𝛼2 𝑆𝑖,2 + 𝛼3 𝑆𝑖,3 + 𝛼4 𝑆𝑖,4 + 𝜀𝑖 (3) Energy consumption intensity of different types of functional sectors can be obtained by this model, and the regression coefficient can also be estimated by least squares method.
-p
3. Model analysis
na
lP
re
3.1. Accuracy verification The area of different functional sectors is firstly determined, and real-time sub-metering data of all the buildings are obtained from energy consumption monitoring platforms in large public buildings. Reference energy consumption intensity of functional sectors (office, shopping mall, hotel and garage) in the trial period can also be obtained. Then the total reference energy consumption can be calculated after adding the reference energy consumption of different functional sectors. In order to study whether this method is suitable for data analysis in energy consumption monitoring platforms, this paper selected 41 large public buildings measured by on-site platforms for initial calculation.
Jo
ur
Fig. 2 shows energy consumption of each sub-item in different buildings from April 7 to April 20 in 2016. The abscissa in the graph refers to the number of different buildings, and the ordinate means energy consumption calculated by equation (1) during every single day. A large number of points in the graph are the electricity reference value of a certain building at different dates. Different areas of functional sectors contributed to the difference in energy consumption among various buildings. Energy consumption of different days in a certain building differed due to weather and work schedule (For example, energy consumption on weekends was significantly lower). From Fig. 2(a), it can be observed that test data during these two weeks basically overlapped, indicating that in many buildings, energy consumption of lighting and sockets in different functional sectors was obviously stable on both weekends and workdays. In some buildings, the points of weekends and workdays were close, which means that the energy consumption of lighting and sockets in commercial and hotel buildings was not much related to workdays. Since in some buildings, the office area accounted for much, the difference of energy consumption on weekends and workdays was a little large.
ro of
a) Lighting and sockets.
Jo
ur
na
lP
re
-p
The electricity of air conditioning, power equipment, and total were also calculated respectively. The results are shown in Fig. 2(b), (c) and (d).
b) Air-conditioning.
ur
ro of
na
lP
re
-p
c) Power equipment.
Jo
Fig. 2. Daily electricity consumption from April
d) Total. 7 th
to 20th.
Since 41 building samples were selected, the error of energy consumption could be much eliminated caused by operation, maintenance, design, and occupants’ flow in certain building. It can be found from the analysis above that reference energy consumption through initial calculation was within a reasonable range. In addition, the difference between weekends and weekdays was logically corresponding with the actual situation. Therefore, this model can be applied into the calculation of reference energy intensity in different functional sectors. This reference energy consumption combines areas in different functional sectors of each building. Therefore, this reference value can be considered as a benchmark for energy consumption evaluation, so as to diagnose whether energy saving could be further carried out.
3.2. Deviation analysis In order to analyze the deviation between energy consumption index of multifunctional buildings and pure functional buildings in linear regression model, this paper selected 20 multifunctional buildings and 30 pure functional buildings (including 10 pure hotel buildings, 12 pure office buildings and 8 pure shopping malls). A set of sub-item energy consumption index applied in both pure functional and multifunctional buildings were calculated based on 50 different types of buildings. All the data were from on-site measurement. The fitting results are shown in Fig. 3 to Fig. 6.
Jo
ur
na
lP
re
-p
ro of
Firstly, the energy consumption index was calculated according to energy consumption intensity and area of pure hotels. And energy intensity values of different sub-items, such as air-conditioning, lighting, and chiller, were obtained. According to equation (3), the energy consumption intensity of different sub-items in various buildings were calculated and compared. As can be seen from Fig. 3, the two groups of data was in high agreement. Table 1 shows the absolute deviation and deviation percentage of energy consumption index for multifunctional and pure functional buildings. It can be observed that the relative error for sub-items was within 4.2% between annual energy consumption intensity of the hotel sector in multifunctional buildings and pure hotels. And the difference of total energy consumption was only 1.1%. Though the relative error of sub-item energy consumption was a little larger, it was also within a reasonable range.
Fig. 3. Comparison of energy consumption index between pure hotels and hotel sector in comprehensive buildings.
It can also be observed from Fig. 4 and Fig. 5 that two sets of data were also in high consistence when pure office buildings and pure shopping malls were compared with other multifunctional buildings. Table 1 shows that the difference of energy consumption index was within 12% between pure shopping malls and shopping mall sector in comprehensive buildings, and the
lP
re
-p
ro of
difference of total energy consumption was 1.1%. Except for heat source and heat pump, the deviations of other sub-items were all less than 3.3%, indicating that heat source system among shopping malls are much different. As for the office sector, the deviation of energy consumption between pure office buildings and office sector in multifunctional buildings ranged from 3.9% to 55.6%, and the difference between the total energy consumption was 3.9%. Except for chilled water pumps and heat pumps, the deviations of the other sub-items ranged from 3.9% to 16.7%. Although certain relative errors of sub-item energy consumption, such as chilled water pumps and heat pumps, appeared large compared with that of pure office buildings, it was related to sub-item energy consumption itself. The absolute deviations of the energy consumption index in chilled water pumps and heat pumps were low, with only 1.1 kWhe/(m2·a) and 1.0 kWhe/(m2·a), respectively.
Jo
ur
buildings.
na
Fig. 4. Comparison of energy consumption index between pure office and office sector in comprehensive
ro of
Fig. 5. Comparison of energy consumption index between pure shopping malls and shopping mall sector in
-p
comprehensive buildings.
Jo
ur
na
lP
re
Fig. 6 illustrates deviation percentage. It can be clearly seen that the data were all in good agreement from total energy consumption to sub-item energy consumption, including air conditioning, lighting, cooling source, heat source and the terminal of air conditioning. The deviation of each sub-item energy consumption index was basically within 15%. And as for hotel and shopping mall sectors, the two sets of data were highly consistent. However, the energy consumption deviations of the office functional sector calculated by multifunctional buildings were relatively large. It can be found from some specific samples that the difference of office buildings may be related to some special functional sector, such as data centers in the office buildings. If the functional sector of office buildings can be further subdivided, the data consistency may be well improved.
ro of
Fig. 6. Deviation percentage of energy consumption index between comprehensive buildings and single
-p
functional buildings.
Chilled
source
source
water pump
0.1
1.3
0.3
0.0
0.0
Table 1. Absolute deviation of energy consumption index between multifunctional buildings and single functional buildings (kWhe/(m2·a)).
Hotel
Absolute deviation
3.0
sector
Deviation percentage
1.1%
Office
Absolute deviation
5.0
sector
Deviation percentage
3.9%
Shopping
Absolute deviation
3.6
mall sector
Deviation percentage
1.1%
Airconditioning 1.5
ur
Lighting
Cooling
re
Total
1.4
lP
Items
Heat
Terminal
pump
4.2%
0.3%
2.7%
1.5%
0.0%
0.0%
5.4
2.5
3.2
2.9
2.8
1.1
1.0
8.5%
11.6%
16.7%
14.3%
14.8%
35.5%
55.6%
1.5
3.0
2.1
1.3
1.0
0.0
0.3
1.2%
2.6%
3.3%
9.3%
2.5%
0.0%
12.0%
4. Conclusions
With a large amount of data accumulated, energy consumption monitoring platforms still lack enough diagnosis and analysis technology. This paper proposes a linear regression model to evaluate building energy consumption based on data obtained from 50 multiple and single function buildings. The following conclusions can be drawn: (1) The energy consumption index of multifunctional buildings calculated by the regression model was consistent with that of pure functional buildings. Therefore, the results of sub-item energy consumption index can also be used to evaluate the energy consumption level in pure functional buildings. In addition, energy consumption intensity of different functional zones in multifunctional buildings can be calculated according to the model. And the energy consumption range can be reasonably determined, which provides a reasonable evaluation method for online energy-saving diagnosis in large-scale public buildings.
Jo
Heating
1.3%
na
Function
(2) The deviation of sub-item energy consumption index was small between the multifunctional buildings and pure functional buildings. The error was the lowest for hotel sectors, totally about 1.1%. The deviation in office sectors was 3.9%, mainly due to other special functional sectors. The error could be decreased if the functional sectors of such buildings can be further subdivided. (3) The sub-item energy consumption intensity of each functional zone in Shanghai was obtained, providing reference for fast building energy consumption prediction in Shanghai.
Conflict of interest The authors declared that we do not have any commercial or associative interest that represents conflict of interest in connection with the work submitted.
ro of
Acknowledgement This work was supported by the National Key R&D Program of China (2017YFC0704206) and National Natural Science Foundation of China (51878405).
References
Brownsword, R.A., Fleming, P.D., Powell, J.C., & Pearsall, N. (2005). Sustainable cities-modelling urban energy
-p
supply and demand. Applied Energy, 82, 167-180.
Chung, W., Hui, Y.V., & Lam, Y.M. (2006). Benchmarking the energy efficiency of commercial buildings. Applied
re
Energy, 83, 1-14.
Eom, J., Kyle, P., Clark, L., Patel, P., & Kim, S. (2012). China’s Building Energy Use: A Long-term Perspective
lP
Based on a Detailed Assessment. Pacific Northwest National Laboratory, PNNL-21073. Retrieved from https://www.researchgate.net/publication/255247855_China's_Building_Energy_Use_A_Long-Term_Perspective_ based_on_a_Detailed_Assessment
na
Hu, K. (2018). Research and application of energy consumption quota of public buildings (Master’s thesis). Hunan University, Hunan, China.
ur
Intergovernmental Panel on Climate Change (IPCC). (2014). IPCC Fifth Assessment Report. Retrieved from http://www.ipcc.ch/report/ar5/
Jo
Killy, A., & Fokaides, P.A. (2015). European smart cities: The role of zero energy building. Sustainable Cities and Society, 15, 86-95.
Kinney, S., Piette, M.A., & Berkeley, E.O.L. (2002). Development of a California commercial building benchmarking database. Lawrence Berkeley National Laboratory, LBNL-50676.
Kontokosta, C.E., & Tull, C. (2017). A data-driven predictive model of city-scale energy use in buildings. Applied Energy, 197, 303-317.
Lam, J.C., Wan, K.W., & Liu, D. (2010). Multiple regression models for energy use in air-conditioned office
buildings in different climates. Energy Conversion and Management, 51, 2692-2697.
Lam, J.C., Wan, K.W., & Yang, L. (2008). Sensitivity analysis and energy conservation measures implications. Energy Conversion and Management, 49, 3170-3177.
Lee, W. (2010). Benchmarking the energy performance for cooling purposes in buildings using a novel index-total performance of energy for cooling purposes. Energy, 35, 50-54.
Li, J. (2008). Method research of energy staple and diagnosis based on separate metering system of building (Master’s thesis). Chongqing University, Chongqing, China.
Matson, N.E., & Piette, M.A. (2005). Review of California and National Methods for Energy Performance
ro of
Benchmarking of Commercial Buildings. Lawrence Berkeley National Laboratory.
Natural Bureau of Statistic of China. (2015). China Statistical Yearbook. Beijing: China Statistics Press.
Sandanayake, M., Zhang, G., & Setunge, S. (2018). A comparative method of air emission impact assessment for
-p
building construction activities. Environmental Impact Assessment Review, 68, 1-9.
Thomas, O., Alan, M., & Roberto, L. (2004). Rating the energy performance of buildings. The International
re
Journal of Low Energy and Sustainable Buildings, Vol.3.
Wang, X. (2010). Studies on key technology of sub-metering in commercial buildings (Doctoral dissertation).
lP
Tsinghua University, Beijing, China.
Wang, Y., & Pan, Y. (2014). Single-factor sensitivity analysis of input parameters in an energy model of pro-to-typical office building. Building Energy Efficiency, 42, 9-14.
na
Zhan, X., Li, W., & Hu, T. (2012). Influence of building enclosure structure on building energy consumption in Chongqing. Building Technology, 43, 554-556.
ur
Zhang, W. (2010). Research in energy monitoring field for public building based on seasonal grey forecast theory (Master’s thesis). Hebei University of Technology, Hebei, China.
Jo
Zhang, X. (2009). Establishing energy consumption quota and reducing energy consumption of public buildings to implement the effective supervision to the government office buildings and large-sale public buildings. Construction Science and Technology, 8, 20–38.
Zhao, J., Zhu, N., & Wu, Y. (2009). The analysis of energy consumption of a commercial building in Tianjin. Energy Policy, 2009, 37, 2092-2097.
Zhu, J., Shen, Y., Song, Z., Zhou, D., Zhang, Z., & Kusiak, A. (2019). Data-driven building load profiling and energy management. Sustainable Cities and Society, 101587.
PII:
S2210-6707(19)31650-6
DOI:
https://doi.org/10.1016/j.scs.2019.101893
Reference:
SCS 101893
To appear in:
Sustainable Cities and Society
Received Date:
10 June 2019
Revised Date:
13 October 2019
Accepted Date:
13 October 2019
Please cite this article as: Zheng Z, Wu J, Lian Z, Zhang H, A Method to Evaluate Building Energy Consumption Based on Energy Use Index of Different Functional Sectors, Sustainable Cities and Society (2019), doi: https://doi.org/10.1016/j.scs.2019.101893
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2019 Published by Elsevier.
A Method to Evaluate Building Energy Consumption Based on Energy Use Index of Different Functional Sectors Zhuling Zheng1,2, Jialin Wu1 , Zhiwei Lian1,*, Huibo Zhang1 1
School of Design, Shanghai Jiao Tong University, Shanghai, China 200240
[email protected] [email protected] [email protected] *Corresponding email: [email protected] 2
ro of
Shanghai Jianke Building Energy Service Co., Ltd, Shanghai, China 200032
Highlights
The energy consumption index of multifunctional buildings calculated by the regression model was consistent with that of pure functional buildings. Therefore, the results of sub-item energy consumption index can also be used to evaluate the energy consumption level in pure functional buildings.
The deviation of sub-item energy consumption index was small between the multifunctional buildings and pure functional buildings.
The sub-item energy intensity of each functional zone in Shanghai was obtained, providing reference for fast building energy consumption prediction in Shanghai.
lP
re
-p
Abstract
Jo
ur
na
Appropriate management of energy consumption provides great support for the sustainable development of buildings and even the whole cities. Nowadays, comprehensive energy consumption monitoring index is widely used in many practical projects to make horizontal comparisons among buildings. However, problems may occur when this method is applied in comprehensive buildings. This paper proposed a method to evaluate the overall energy consumption of buildings based on the energy index obtained from different functional sectors. A linear regression model was developed to predict energy consumption index in multifunctional areas. The model was based on online monitoring data from 30 single functional buildings and 20 multifunctional buildings. The sub-item energy consumption intensity difference between pure hotel buildings and hotel sector in comprehensive buildings was within 4.2%. The deviation of sub-item energy consumption intensity was less than 12% between pure shopping malls and shopping mall sector in comprehensive buildings. As for the pure office buildings and office sector in comprehensive buildings, except for heat pumps and chilled water pumps, the deviations of energy intensity of other sub-items ranged from 3.9% to 16.7%. The results from on-site experiments show that the sub-item energy consumption index obtained from multifunctional buildings had lower error than that of single functional buildings. Therefore, the sub-item energy consumption index can also be used to evaluate energy consumption in pure functional buildings. The model proposed in this paper provides a good reference for the prediction of energy consumption index in large public buildings.
Keywords Energy consumption; Linear regression; Single functional buildings; Multifunctional buildings
1. Introduction
ro of
China has now become the second largest country in building energy consumption (Eom, Kyle, Clark, Patel & Kim, 2012). The energy consumption in public buildings accounted for about 26.4%, and the energy consumption per unit area increased from 16.5kgce/m2 in 2001 to 21.9kgce/m2 in 2012 at an average rate of 7% annually (IPCC Fifth Assessment Report, 2014). According to the statistics about the electricity consumption intensity of public buildings in China in recent years (China Statistical Yearbook, 2015), the growth of public building area slowed down, while electricity consumption intensity increased exponentially. Urban construction greatly created an environmental burden, triggering the acceleration of climate change (Sandanayake, Zhang & Setunge, 2018). In the construction process, it is much limited that new materials, equipment, systems and other technologies could be adopted to reduce energy consumption. Actually, appropriate management of energy consumption may also provide great support for sustainable development of buildings and even the whole cities. Therefore, intelligent energy management recently became the main source of energy conservation (Killy & Fokaides, 2015).
Jo
ur
na
lP
re
-p
However, with a large amount of online monitoring data obtained from energy consumption monitoring platforms, it is necessary to provide online energy-saving diagnosis techniques. Kontokosta and Tull (2017) used several algorithms, including linear fit, random forest, and support vector machine, to build a data-driven model based on the energy consumption data from 23,000 buildings and predicted the energy consumption of 1.1 million buildings in New York. Brownsword, Fleming, Powell and Pearsall (2005) established a fitting algorithm for the energy demand of different consumption types in Leicester according to survey data, and then the algorithm was used to predict short-term energy demand. Zhu et al. (2019) presented a systematic way of qualifying daily building load patterns and identifying abnormal energy consumption. These were real-time energy-saving diagnosis based on the energy consumption prediction. However, the algorithm itself was much complicated and the accuracy of the algorithm strongly depended on the model and parameter setting, which required enough historical data or a large number of building samples. Wang (2010) proposed three basic energy-saving diagnostic methods, including efficiency indicator method, and provided specific diagnosis cases. Zhang (2010) proposed a seasonal gray theory prediction algorithm by improving the smoothness and residual model, which was applied in energy conservation monitoring platforms of public buildings in Shanxi, China. Although these algorithms were simplified and applied into practice, only certain types of buildings were included and the monitoring platforms still cannot serve for most common types of buildings. Some scholars have studied the energy consumption quota in various types of buildings. In some research (Lee, 2010; Kinney, Piette & Berkeley, 2002; Thomas, Alan & Roberto, 2004), the evaluation methods of the total energy consumption in several buildings were given. EUI (Energy Use intensity) index or normalized EUI index was adopted to measure the overall energy use of the building. Chung, Hui and Lam (2006) introduced EUI index and found its influencing factors based on multiple regression analysis, and a benchmark electricity meter for EUI was obtained. In terms of Cal-Arch, one of the energy evaluation tools, the distribution of energy consumption in various buildings was explored from the CEUS database in the United States and Canada, and the
EUI index of each building was compared according to the distribution histogram to evaluate the building energy (Matson & Piette, 2005). Zhang (2009) attempted to find an energy consumption quota where 60% of building energy consumption was guaranteed to be less than this value, according to energy consumption survey and information collection. Hu (2018) obtained a reasonable quota level by regression analysis and energy consumption quotas, which ensured that energy consumption in most buildings is lower than the control level, providing a good basis for energy conservation supervision. However, several problems existed in this quota level. For example, overall energy consumption could meet the standard while the local part may exceed the standard. If the energy consumption among buildings cannot be objectively evaluated, energy-saving diagnosis and renovation in large-scale public buildings will be affected.
-p
ro of
Li (2008) proposed crossing method on sub-metering system platforms. Local diagnosis was first carried out and energy-saving potential was then comprehensively explored in public buildings. Zhao, Zhu and Wu (2009) analyzed the energy consumption of public buildings in detail based on energy-saving behavior, thermal environment and air-conditioning operation management, and methods for energy consumption diagnosis were also given. These strategies were used to save energy mainly from on-site operations of the equipment. Therefore, this paper proposes a regression model based on massive sub-item energy consumption data. The model also provides new thoughts for online energy-saving diagnosis.
re
2. Methods
Jo
ur
na
lP
2.1. Distribution of building energy consumption In order to find out the distribution of energy consumption in different functional buildings, this paper collected enough energy consumption data from 130 public buildings in Shanghai. The energy consumption intensity of each building is shown in Fig. 1.
Fig. 1. Total energy consumption intensity of 130 large public buildings.
According to the statistics, the total energy consumption per unit area ranged from 50.5kWh/(m2 · a) to 430.2kWh/(m2·a), and the difference was obvious. The total energy
ro of
consumption of shopping malls and supermarkets was the highest, with the average value of 271.6kWh/(m2·a) and 281.7kWh/(m2·a) respectively, and that of commercial and hotel buildings was 147.9kWh/(m2·a) and 142.7kWh/(m2·a) respectively. Energy consumption in office buildings was relatively low, with the average value of 107kWh/(m2·a). However, comprehensive commercial buildings occupy a large part of cities. Generally, shopping mall sectors are arranged at the lower floor of comprehensive buildings and the upper part is office sectors. And the disparity of energy consumption per unit area is large between shopping mall and office sectors. Therefore, the proportion of various functional areas is different, resulting in the difference of energy intensity among types of multifunctional buildings. Reasonable evaluation method about building energy consumption should be introduced. It can be found that even if the buildings include the same functional sectors, the area of these sectors will greatly affect the overall building energy consumption. Thus, this paper intends to establish an evaluation model based on the functional area.
re
-p
2.2. Regression model Some scholars have used multiple linear regression methods to study the sensitivity of design parameters, such as envelope, HVAC system and building materials, towards building energy consumption (Zhan, Li & Hu, 2012; Wang & Pan, 2014; Lam, Wan & Yang, 2008; Lam, Wan & Liu, 2010). Since this paper considers the impact of functional area on the overall energy consumption, each functional sector can be regarded as the “weight” affecting energy consumption. Therefore, linear regression model is also adopted in this paper for data fitting.
lP
It is assumed that yi be total energy consumption, where i=1, 2...n (n is the number of samples). Since the area of functional sectors greatly affects the overall energy consumption, it can be considered as an independent variable in the regression model. Si,j means the area of different functional sectors, where 𝑗 = 1, … , m (m is the type of functional sectors).
ur
na
Generally, functional sectors can be divided into the following seven categories: office, accommodation, shopping, sports, education, medical treatment and others. Four types of most common functional sectors are involved in this paper: office, hotel, shopping mall, garage, which are also covered by the selected building samples. Therefore, m is taken as 4.
Jo
In order to find the relationship between functional area and total energy consumption and obtain the basic parameters, general linear regression model is proposed as equation (1). 𝑦𝑖 = 𝛼0 + 𝛼1 𝑆𝑖,1 + 𝛼2 𝑆𝑖,2 + 𝛼3 𝑆𝑖,3 + 𝛼4 𝑆𝑖,4 + 𝜀𝑖 (1) Where, yi——sub-item energy consumption, kWh/d; α1 ——sub-item energy consumption intensity of hotel sector, kWh/(m2·d); α2——sub-item energy consumption intensity of office sector, kWh/(m2·d); α3——sub-item energy consumption intensity of shopping mall sector, kWh/(m2·d); α4——sub-item energy consumption intensity of garage sector, kWh/(m2·d); Si,1——area of hotel sector, m2; Si,2——area of office sector, m2; Si,3——area of shopping mall sector, m2; Si,4——area of garage sector, m2;
εi——error, kWh/d. Energy consumption intensity is an important index to evaluate energy consumption in certain functional sector. Sub-item energy consumption intensity in certain functional sector is defined as equation (2):
j=
W
j
(2)
Sj
Where, αj——sub-item energy consumption intensity, kWh/m2; Wj——sub-item energy consumption in certain functional sector, kWh; Sj——area of certain functional sector, m2.
ro of
Since the total energy consumption involves each functional sector, it is assumed that 𝛼0 = 0. Equation (1) can be further simplified to equation (3): 𝑦𝑖 = 𝛼1 𝑆𝑖,1 + 𝛼2 𝑆𝑖,2 + 𝛼3 𝑆𝑖,3 + 𝛼4 𝑆𝑖,4 + 𝜀𝑖 (3) Energy consumption intensity of different types of functional sectors can be obtained by this model, and the regression coefficient can also be estimated by least squares method.
-p
3. Model analysis
na
lP
re
3.1. Accuracy verification The area of different functional sectors is firstly determined, and real-time sub-metering data of all the buildings are obtained from energy consumption monitoring platforms in large public buildings. Reference energy consumption intensity of functional sectors (office, shopping mall, hotel and garage) in the trial period can also be obtained. Then the total reference energy consumption can be calculated after adding the reference energy consumption of different functional sectors. In order to study whether this method is suitable for data analysis in energy consumption monitoring platforms, this paper selected 41 large public buildings measured by on-site platforms for initial calculation.
Jo
ur
Fig. 2 shows energy consumption of each sub-item in different buildings from April 7 to April 20 in 2016. The abscissa in the graph refers to the number of different buildings, and the ordinate means energy consumption calculated by equation (1) during every single day. A large number of points in the graph are the electricity reference value of a certain building at different dates. Different areas of functional sectors contributed to the difference in energy consumption among various buildings. Energy consumption of different days in a certain building differed due to weather and work schedule (For example, energy consumption on weekends was significantly lower). From Fig. 2(a), it can be observed that test data during these two weeks basically overlapped, indicating that in many buildings, energy consumption of lighting and sockets in different functional sectors was obviously stable on both weekends and workdays. In some buildings, the points of weekends and workdays were close, which means that the energy consumption of lighting and sockets in commercial and hotel buildings was not much related to workdays. Since in some buildings, the office area accounted for much, the difference of energy consumption on weekends and workdays was a little large.
ro of
a) Lighting and sockets.
Jo
ur
na
lP
re
-p
The electricity of air conditioning, power equipment, and total were also calculated respectively. The results are shown in Fig. 2(b), (c) and (d).
b) Air-conditioning.
ur
ro of
na
lP
re
-p
c) Power equipment.
Jo
Fig. 2. Daily electricity consumption from April
d) Total. 7 th
to 20th.
Since 41 building samples were selected, the error of energy consumption could be much eliminated caused by operation, maintenance, design, and occupants’ flow in certain building. It can be found from the analysis above that reference energy consumption through initial calculation was within a reasonable range. In addition, the difference between weekends and weekdays was logically corresponding with the actual situation. Therefore, this model can be applied into the calculation of reference energy intensity in different functional sectors. This reference energy consumption combines areas in different functional sectors of each building. Therefore, this reference value can be considered as a benchmark for energy consumption evaluation, so as to diagnose whether energy saving could be further carried out.
3.2. Deviation analysis In order to analyze the deviation between energy consumption index of multifunctional buildings and pure functional buildings in linear regression model, this paper selected 20 multifunctional buildings and 30 pure functional buildings (including 10 pure hotel buildings, 12 pure office buildings and 8 pure shopping malls). A set of sub-item energy consumption index applied in both pure functional and multifunctional buildings were calculated based on 50 different types of buildings. All the data were from on-site measurement. The fitting results are shown in Fig. 3 to Fig. 6.
Jo
ur
na
lP
re
-p
ro of
Firstly, the energy consumption index was calculated according to energy consumption intensity and area of pure hotels. And energy intensity values of different sub-items, such as air-conditioning, lighting, and chiller, were obtained. According to equation (3), the energy consumption intensity of different sub-items in various buildings were calculated and compared. As can be seen from Fig. 3, the two groups of data was in high agreement. Table 1 shows the absolute deviation and deviation percentage of energy consumption index for multifunctional and pure functional buildings. It can be observed that the relative error for sub-items was within 4.2% between annual energy consumption intensity of the hotel sector in multifunctional buildings and pure hotels. And the difference of total energy consumption was only 1.1%. Though the relative error of sub-item energy consumption was a little larger, it was also within a reasonable range.
Fig. 3. Comparison of energy consumption index between pure hotels and hotel sector in comprehensive buildings.
It can also be observed from Fig. 4 and Fig. 5 that two sets of data were also in high consistence when pure office buildings and pure shopping malls were compared with other multifunctional buildings. Table 1 shows that the difference of energy consumption index was within 12% between pure shopping malls and shopping mall sector in comprehensive buildings, and the
lP
re
-p
ro of
difference of total energy consumption was 1.1%. Except for heat source and heat pump, the deviations of other sub-items were all less than 3.3%, indicating that heat source system among shopping malls are much different. As for the office sector, the deviation of energy consumption between pure office buildings and office sector in multifunctional buildings ranged from 3.9% to 55.6%, and the difference between the total energy consumption was 3.9%. Except for chilled water pumps and heat pumps, the deviations of the other sub-items ranged from 3.9% to 16.7%. Although certain relative errors of sub-item energy consumption, such as chilled water pumps and heat pumps, appeared large compared with that of pure office buildings, it was related to sub-item energy consumption itself. The absolute deviations of the energy consumption index in chilled water pumps and heat pumps were low, with only 1.1 kWhe/(m2·a) and 1.0 kWhe/(m2·a), respectively.
Jo
ur
buildings.
na
Fig. 4. Comparison of energy consumption index between pure office and office sector in comprehensive
ro of
Fig. 5. Comparison of energy consumption index between pure shopping malls and shopping mall sector in
-p
comprehensive buildings.
Jo
ur
na
lP
re
Fig. 6 illustrates deviation percentage. It can be clearly seen that the data were all in good agreement from total energy consumption to sub-item energy consumption, including air conditioning, lighting, cooling source, heat source and the terminal of air conditioning. The deviation of each sub-item energy consumption index was basically within 15%. And as for hotel and shopping mall sectors, the two sets of data were highly consistent. However, the energy consumption deviations of the office functional sector calculated by multifunctional buildings were relatively large. It can be found from some specific samples that the difference of office buildings may be related to some special functional sector, such as data centers in the office buildings. If the functional sector of office buildings can be further subdivided, the data consistency may be well improved.
ro of
Fig. 6. Deviation percentage of energy consumption index between comprehensive buildings and single
-p
functional buildings.
Chilled
source
source
water pump
0.1
1.3
0.3
0.0
0.0
Table 1. Absolute deviation of energy consumption index between multifunctional buildings and single functional buildings (kWhe/(m2·a)).
Hotel
Absolute deviation
3.0
sector
Deviation percentage
1.1%
Office
Absolute deviation
5.0
sector
Deviation percentage
3.9%
Shopping
Absolute deviation
3.6
mall sector
Deviation percentage
1.1%
Airconditioning 1.5
ur
Lighting
Cooling
re
Total
1.4
lP
Items
Heat
Terminal
pump
4.2%
0.3%
2.7%
1.5%
0.0%
0.0%
5.4
2.5
3.2
2.9
2.8
1.1
1.0
8.5%
11.6%
16.7%
14.3%
14.8%
35.5%
55.6%
1.5
3.0
2.1
1.3
1.0
0.0
0.3
1.2%
2.6%
3.3%
9.3%
2.5%
0.0%
12.0%
4. Conclusions
With a large amount of data accumulated, energy consumption monitoring platforms still lack enough diagnosis and analysis technology. This paper proposes a linear regression model to evaluate building energy consumption based on data obtained from 50 multiple and single function buildings. The following conclusions can be drawn: (1) The energy consumption index of multifunctional buildings calculated by the regression model was consistent with that of pure functional buildings. Therefore, the results of sub-item energy consumption index can also be used to evaluate the energy consumption level in pure functional buildings. In addition, energy consumption intensity of different functional zones in multifunctional buildings can be calculated according to the model. And the energy consumption range can be reasonably determined, which provides a reasonable evaluation method for online energy-saving diagnosis in large-scale public buildings.
Jo
Heating
1.3%
na
Function
(2) The deviation of sub-item energy consumption index was small between the multifunctional buildings and pure functional buildings. The error was the lowest for hotel sectors, totally about 1.1%. The deviation in office sectors was 3.9%, mainly due to other special functional sectors. The error could be decreased if the functional sectors of such buildings can be further subdivided. (3) The sub-item energy consumption intensity of each functional zone in Shanghai was obtained, providing reference for fast building energy consumption prediction in Shanghai.
Conflict of interest The authors declared that we do not have any commercial or associative interest that represents conflict of interest in connection with the work submitted.
ro of
Acknowledgement This work was supported by the National Key R&D Program of China (2017YFC0704206) and National Natural Science Foundation of China (51878405).
References
Brownsword, R.A., Fleming, P.D., Powell, J.C., & Pearsall, N. (2005). Sustainable cities-modelling urban energy
-p
supply and demand. Applied Energy, 82, 167-180.
Chung, W., Hui, Y.V., & Lam, Y.M. (2006). Benchmarking the energy efficiency of commercial buildings. Applied
re
Energy, 83, 1-14.
Eom, J., Kyle, P., Clark, L., Patel, P., & Kim, S. (2012). China’s Building Energy Use: A Long-term Perspective
lP
Based on a Detailed Assessment. Pacific Northwest National Laboratory, PNNL-21073. Retrieved from https://www.researchgate.net/publication/255247855_China's_Building_Energy_Use_A_Long-Term_Perspective_ based_on_a_Detailed_Assessment
na
Hu, K. (2018). Research and application of energy consumption quota of public buildings (Master’s thesis). Hunan University, Hunan, China.
ur
Intergovernmental Panel on Climate Change (IPCC). (2014). IPCC Fifth Assessment Report. Retrieved from http://www.ipcc.ch/report/ar5/
Jo
Killy, A., & Fokaides, P.A. (2015). European smart cities: The role of zero energy building. Sustainable Cities and Society, 15, 86-95.
Kinney, S., Piette, M.A., & Berkeley, E.O.L. (2002). Development of a California commercial building benchmarking database. Lawrence Berkeley National Laboratory, LBNL-50676.
Kontokosta, C.E., & Tull, C. (2017). A data-driven predictive model of city-scale energy use in buildings. Applied Energy, 197, 303-317.
Lam, J.C., Wan, K.W., & Liu, D. (2010). Multiple regression models for energy use in air-conditioned office
buildings in different climates. Energy Conversion and Management, 51, 2692-2697.
Lam, J.C., Wan, K.W., & Yang, L. (2008). Sensitivity analysis and energy conservation measures implications. Energy Conversion and Management, 49, 3170-3177.
Lee, W. (2010). Benchmarking the energy performance for cooling purposes in buildings using a novel index-total performance of energy for cooling purposes. Energy, 35, 50-54.
Li, J. (2008). Method research of energy staple and diagnosis based on separate metering system of building (Master’s thesis). Chongqing University, Chongqing, China.
Matson, N.E., & Piette, M.A. (2005). Review of California and National Methods for Energy Performance
ro of
Benchmarking of Commercial Buildings. Lawrence Berkeley National Laboratory.
Natural Bureau of Statistic of China. (2015). China Statistical Yearbook. Beijing: China Statistics Press.
Sandanayake, M., Zhang, G., & Setunge, S. (2018). A comparative method of air emission impact assessment for
-p
building construction activities. Environmental Impact Assessment Review, 68, 1-9.
Thomas, O., Alan, M., & Roberto, L. (2004). Rating the energy performance of buildings. The International
re
Journal of Low Energy and Sustainable Buildings, Vol.3.
Wang, X. (2010). Studies on key technology of sub-metering in commercial buildings (Doctoral dissertation).
lP
Tsinghua University, Beijing, China.
Wang, Y., & Pan, Y. (2014). Single-factor sensitivity analysis of input parameters in an energy model of pro-to-typical office building. Building Energy Efficiency, 42, 9-14.
na
Zhan, X., Li, W., & Hu, T. (2012). Influence of building enclosure structure on building energy consumption in Chongqing. Building Technology, 43, 554-556.
ur
Zhang, W. (2010). Research in energy monitoring field for public building based on seasonal grey forecast theory (Master’s thesis). Hebei University of Technology, Hebei, China.
Jo
Zhang, X. (2009). Establishing energy consumption quota and reducing energy consumption of public buildings to implement the effective supervision to the government office buildings and large-sale public buildings. Construction Science and Technology, 8, 20–38.
Zhao, J., Zhu, N., & Wu, Y. (2009). The analysis of energy consumption of a commercial building in Tianjin. Energy Policy, 2009, 37, 2092-2097.
Zhu, J., Shen, Y., Song, Z., Zhou, D., Zhang, Z., & Kusiak, A. (2019). Data-driven building load profiling and energy management. Sustainable Cities and Society, 101587.