Online Methodology for Separating the Power Consumption of Lighting Sockets and Air-Conditioning in Public Buildings Based on an Outdoor Temperature Partition Model and Historical Energy Consumption Data
Abstract
:1. Introduction
2. Background and Data for Power Monitoring
3. Pre-Processing of Power Consumption Data
- The K sets of points are randomly selected as the initial clustering center in the sample data set. According to the three kinds of air-conditioning state (closed/slightly open/fully open) and three kinds of data state (located/above/below), K is evaluated at 3 for KNN throughout the paper.
- The distance between other points and the initial clustering center point is calculated, and these other points are allocated to the nearest neighbor cluster.
- After the preliminary clustering is completed, the averages of all sample points in different clusters are selected as the new clustering center, and then steps 1 and 2 are repeated.
- The clustering center and clusters of the sample points are updated iteratively, until the clustering center no longer changes, which means the end of this clustering algorithm. Next, we can output the clustering center and K pieces of the clusters of the sample points.
4. Separation Methodology
- Use the model of average outdoor temperature and lighting socket daily power consumption, or the k-means clustering algorithm, to obtain the eigenvalues of the method, including clustering center (c), outdoor temperature threshold (), and power consumption data threshold (), etc.
- Use the above eigenvalues and the lighting socket power consumption with air-conditioning opened to identify the state of the air-conditioning.
- Use appropriate historical data of lighting sockets (without mixing the power consumption of the air-conditioning) to predict and interpolate the actual power consumption of the lighting socket when the air-conditioning is running.
4.1. Method for Identifying Abnormal Conditions for Daily Data
4.2. The Case of Identifying Abnormal Conditions for Daily Data
4.3. Method for Identifying Abnormal Conditions for Hourly Data
- (1)
- Using the clustering algorithm to generate the critical eigenvalues, and , when the power consumption and the outdoor temperature are less than and , respectively, the state of the air-conditioning is closed. For the contrary state, the air-conditioning is open.
- (2)
- The theoretical maximum power consumption of the lighting socket in this special hour is calculated and compared with the actual power consumption . If is higher than , it is considered that the power consumption of the lighting socket is mixed with that of the air conditioner; otherwise, it is not mixed.
4.4. The Cases of Identifying Abnormal Conditions for Hourly Data
4.5. Comparison of the OATPM Method and the KNN Method
4.6. The Method and Case of Separating Abnormal Data after Identifying Data
5. Discussion of the Separation Methodology
6. Selection of Important Parameters in the Clustering Algorithm
7. Conclusions
- (1)
- The OATPM and KNN methods are driven by the historical data of the energy monitoring platform, which can effectively separate the power consumption of lighting sockets and air-conditioning in public buildings. According to the three kinds of air-conditioning state (closed/slightly open/fully open) and three kinds of data state (located/above/below), K is evaluated at 3 for KNN throughout the paper. The identification error for three public buildings utilizing the method was less than 15%, and the proportion of error greater than 10% was less than 15%.
- (2)
- The OATPM method is suitable for identifying and separating daily data, and the calculation speed is high, but it cannot identify and separate hourly data. Thus, it is suitable for scenes with low identification accuracy, such as research on total energy consumption statistics. The KNN method is suitable for identifying and separating not only daily data but also hourly data; however, the calculation is complex and slow. Therefore, it is suitable for scenes with high identification accuracy, such as research on the correlation between human behavior characteristics and energy consumption.
- (3)
- The methodology proposed in this study is suitable for public buildings with different functions and climates, especially for buildings with high power consumption values for lighting sockets or large differences between power consumption characteristics of air-conditioning and lighting sockets. For example, in commercial buildings, the power consumption of air-conditioning is relatively large. Therefore, the power consumption of lighting sockets is far less than the power consumption of lighting sockets (mixed air-conditioning) in the cooling season, which means that the number of clustering iterations is fewer, and the distances between different clusters is greater. Thus, the identification results of this separation methodology for commercial buildings are accurate.
Author Contributions
Funding
Institutional Review Board Statement
Informed Consent Statement
Conflicts of Interest
References
- China Association of Building Energy Efficiency. Report on Chinese Building Energy Consumption in 2018. Available online: http://www.cabee.org/site/content/22960.html (accessed on 9 December 2018).
- China Association of Building Energy Conservation, Energy Consumption Statistics Committee of China Association of Building Energy Conservation. Report of China Building Energy Consumption Research. Available online: https://www.cabee.org/site/content/23565.html (accessed on 10 April 2020).
- Deng, S.; Burnett, J. A study of energy performance of hotel buildings in Hong Kong. Energy Build. 2000, 31, 7–12. [Google Scholar] [CrossRef]
- Lee, W.L.; Yik, F.W.H.; Jones, P.; Burnett, J. Energy saving by realistic design data for commercial buildings in Hong Kong. Appl. Energy 2001, 70, 59–75. [Google Scholar] [CrossRef]
- Philip, C.H.; Chow, W.K. Energy use in commercial buildings in Hong Kong. Energy 2001, 69, 243–255. [Google Scholar]
- Benavente-Peces, C.; Ibadah, N. Buildings Energy Efficiency Analysis and Classification Using Various Machine Learning Technique Classifiers. Energies 2020, 13, 3497. [Google Scholar] [CrossRef]
- Monteiro, S.A.; Monteiro, F.P.; Tostes, M.E.L.; Carvalho, C.M. Methodology for Energy Efficiency on Lighting and Air Conditioning Systems in Buildings Using a Multi-Objective Optimization Algorithm. Energies 2020, 13, 3303. [Google Scholar] [CrossRef]
- Zhang, Y.; Bai, X.; Mills, F.P.; Pezzey, J.C.V. Rethinking the role of occupant behavior in building energy performance: A review. Energy Build. 2018, 172, 279–294. [Google Scholar] [CrossRef]
- Li, Q.; Sun, X.; Chen, C.; Yang, X. Characterizing the household energy consumption in heritage Nanjing Tulou buildings, China: A comparative field survey study. Energy Build. 2012, 49, 317–326. [Google Scholar] [CrossRef]
- Hoes, P.; Hensen, J.L.M.; Loomans, M.G.L.C.; de Vries, B.; Bourgeois, D. User behavior in whole building simulation. Energy Build. 2009, 41, 295–302. [Google Scholar] [CrossRef] [Green Version]
- Ioannou, A.; Itard, L.C. Energy performance and comfort in residential buildings: Sensitivity for building parameters and occupancy. Energy Build. 2015, 92, 216–233. [Google Scholar] [CrossRef]
- Wang, C. Simulation Research on Occupant Energy-Related Behaviors in Building; Tsinghua University: Beijing, China, November 2014. [Google Scholar]
- Zhou, G.; Han, Z.; Fu, J.; Xu, G.H.; Ye, C. Iterative Online Fault Identification Scheme for High-Voltage Circuit Breaker Utilizing a Lost Data Repair Technique. Energies 2020, 13, 3311. [Google Scholar] [CrossRef]
- Ministry of House and Urban-Rural Development of People’s Republic of China. Implementation Opinions on Strengthening the Administration of Energy Conservation in State Organs Office Buildings and Large Public Buildings. 2007. Available online: http://www.mohurd.gov.cn/wjfb/200710/t20071026_158566.html (accessed on 23 October 2007).
- Ministry of House and Urban-Rural Development of People’s Republic of China. Interim Measures for the Acceptance and Operation Management of Provincial-Level Public Building Energy Consumption Monitoring Platform. 2016. Available online: http://www.mohurd.gov.cn/wjfb/201604/t20160415_227217.html (accessed on 11 April 2016).
- Marceau, M.L.; Zmeureanu, R. Nonintrusive load disaggregation computer program to estimate the energy consumption of major end uses in residential buildings. Energy Convers. Manag. 2000, 41, 1389–1403. [Google Scholar] [CrossRef] [Green Version]
- Pihala, H. Non-intrusive Appliance Load Monitoring System Based on a Modern kWh- Meter; VTT Publications: Espoo, Finland, 1998; p. 356. [Google Scholar]
- Norford, L.K.; Leeb, S.B. Non-intrusive electrical load monitoring in commercial buildings based on steady-state and transient load-detection algorithms. Energy Build. 1996, 24, 51–64. [Google Scholar] [CrossRef]
- Akbari, H.; Konopacki, S.J. Application of an End-Use Disaggregation Algorithm for Obtaining Building Energy-Use Data. J. Sol. Energy Eng. 1998, 120, 205–210. [Google Scholar] [CrossRef]
- Li, X.; Bowers, C.P.; Schnier, T. Classification of Energy Consumption in Buildings with Outlier Detection. IEEE Trans. Ind. Electron. 2010, 57, 3639–3644. [Google Scholar] [CrossRef]
- Wang, X. Studies on Key Technology of Sub-Metering in Commercial Buildings; Tsinghua University: Beijing, China, November 2010. [Google Scholar]
- Doherty, B.; Trenbath, K. Device-level plug load disaggregation in a zero energy office building and opportunities for energy savings. Energy Build. 2019, 204, 109480. [Google Scholar] [CrossRef]
- Anand, P.; Cheong, D.; Sekhar, C.; Santamouris, M.; Kondepudi, S. Energy saving estimation for plug and lighting load using occupancy analysis. Renew. Energy 2019, 143, 1143–1161. [Google Scholar] [CrossRef]
- Aiad, M.; Lee, P.H. Energy disaggregation of overlapping home appliances consumptions using a cluster separating approach. Sustain. Cities Soc. 2018, 44, 487–494. [Google Scholar] [CrossRef]
- Aiad, M.; Lee, P.H. Unsupervised approach for load disaggregation with devices interactions. Energy Build. 2016, 116, 96–103. [Google Scholar] [CrossRef]
- Aiad, M.; Lee, P.H. Non-intrusive load disaggregation with adaptive estimations of devices main power effects and two-way interactions. Energy Build. 2016, 130, 131–139. [Google Scholar] [CrossRef]
- Farinaccio, L.; Zmeureanu, R. Using a pattern recognition approach to disaggregate the total electricity consumption in a house into the major end-uses. Energy Build. 1999, 30, 245–259. [Google Scholar] [CrossRef]
- Gabaldón, A.; Molina, R.; Marín-Parra, A.; Valero-Verdu, S.; Alvarez, C. Residential end-uses disaggregation and demand response evaluation using integral transforms. J. Mod. Power Syst. Clean Energy 2017, 5, 91–104. [Google Scholar] [CrossRef] [Green Version]
- Kolter, J.; Jaakkola, T. Approximate inference in additive factorial HMMs with application to energy disaggregation. In Proceedings of the International Conference on Artificial Intelligence and Statistics (AISTATS), La Palma, Canary Islands, Spain, 21–23 April 2012; pp. 1472–1482. [Google Scholar]
- Zhou, Y.; Shi, Z.; Shi, Z.; Gao, Q.; Wu, L. Disaggregating power consumption of commercial buildings based on the finite mixture model. Appl. Energy 2019, 243, 35–46. [Google Scholar] [CrossRef]
- Leonard, K.; Peter, J.R. Finding Groups in Data: An Introduction to Cluster Analysis; John Wiley & Sons: Hoboken, NJ, USA, 1990. [Google Scholar]
- Forgy, E. Cluster analysis of multivariate data: Efficiency vs. interpretability of classifications. Biometrics 1965, 21, 768. [Google Scholar]
- Jain, A.K. Data clustering: 50 years beyond K-means. Pattern Recogn. Lett. 2010, 31, 651–666. [Google Scholar] [CrossRef]
- Steinley, D. K-Means clustering: A half-century synthesis. Brit. J. Math. Stat. Psychol. 2006, 59, 1–34. [Google Scholar] [CrossRef] [PubMed] [Green Version]
- ASHRAE Standards Committee. Measurement of Energy and Demand Savings: ASHRAE Guidelines 14—2002; ASHRAE Inc.: Atlanta, GA, USA, 2002; p. 17. [Google Scholar]
- US Department of Energy Office of Energy Efficiency and Renewable Energy. M&V guidelines: Measurement and verification for federal energy projects. Islington 2008. [Google Scholar] [CrossRef]
Building Codes | Floor Height (m) | Floor Area (m2) | Energy Intensity (kWh/m2·Year) | Climate Partition | Heating Scheme | Cooling Scheme |
---|---|---|---|---|---|---|
Building A | 11.8 | 37,593 | 13.9 | Cold region | Municipal water | Separate air-conditioning |
Building B | 56 | 26,000 | 46.6 | Severely cold region | Municipal water | Separate air-conditioning |
Building C | 41 | 100,000 | 93.9 | Severely cold region | Municipal water | Fan coil unit |
Energy Consumption Monitoring Platform | Building Sample | Total Number of Buildings (Blocks) | Floor Area of Buildings (×1000 m2) | Monitoring Point (Pieces) |
---|---|---|---|---|
Dalian public institution energy consumption monitoring | A | 15 | 300 | 951 |
Liaoning public institution building energy consumption monitoring platform | B, C | 50 | 2030 | 1659 |
State | Workdays | Holidays | ||||
---|---|---|---|---|---|---|
The number of the cluster | 1 | 2 | 3 | 1 | 2 | 3 |
Cluster center (kWh) | 1848.65 | 1263.52 | 485.15 | 1037.01 | 778.52 | 151.75 |
Sample numbers (pieces) | 8 | 130 | 7 | 21 | 51 | 3 |
Proportion (%) | 4.88 | 90.85 | 4.27 | 28.00 | 68.00 | 4.00 |
State conclusion | Abnormal | Reasonable | Abnormal | Reasonable | Reasonable | Abnormal |
Method | Identification Accuracy | Applicable Occasions | Degree of Identification | Can It Be Used to Predict Energy Consumption? |
---|---|---|---|---|
Temperature model | 83.9% | For daily data | Closed/open | Yes |
Clustering algorithm | 93.1% 96.2% | For daily data For hourly data | Closed/slightly open/fully open | No |
Cluster Categories | Number of Samples | Eigenvalues | Power Consumption of Lighting (kWh) | Outdoor Temperature (°C) |
---|---|---|---|---|
Cluster Ⅰ: closed | 47 | C | 1152.50 | 17.28 |
MAX | 1282.65 | 24.55 | ||
MIN | 805.47 | 3.30 | ||
Cluster Ⅱ: slightly open | 27 | C | 1448.11 | 19.21 |
MAX | 1686.75 | 24.83 | ||
MIN | 1287.33 | 19.89 | ||
Cluster Ⅲ: fully open | 13 | C | 2066.55 | 28.87 |
MAX | 2280.90 | 32.05 | ||
MIN | 1787.04 | 25.38 |
Building | Climate Zone | Building Function | Working Time | Form of Air-Conditioning | Accuracy | |
---|---|---|---|---|---|---|
For Daily Data (%) | For Hourly Data (%) | |||||
Building A | Cold | Office building | 7:00–22:00 | Separate air-conditioning | 93.1 | 96.3 |
Building B | Severely cold | Office building | 7:00–17:00 | Separate air-conditioning | 91.2 | 93.4 |
Building C | Severely cold | Commercial building | 8:00–21:00 | Fan coil unit | 95.6 | 96.0 |
Technical Code in China (JGJ176-2009) | IPMVP | FEMP | |
---|---|---|---|
Error in one month | ±15% | ±20% | ±15% |
Error in one year | No Standard | No Standard | ±10% |
Cluster Categories | Number of Samples | Eigenvalues | Power Consumption of Lighting (kWh) | Outdoor Temperature (°C) |
---|---|---|---|---|
Cluster Ⅰ: closed | 124 | C | 1182.52 | 16.00 |
MAX | 1329.90 | 24.95 | ||
MIN | 805.47 | −8.72 | ||
Cluster Ⅱ: slightly open | 74 | C | 1517.24 | 2.52 |
MAX | 1787.04 | 30.47 | ||
MIN | 1353.89 | −12.10 | ||
Cluster Ⅲ: fully open | 25 | C | 2089.84 | 29.09 |
MAX | 2280.90 | 32.05 | ||
MIN | 1610.49 | 25.38 |
Cluster Categories | Number of Samples | Eigenvalues | Power Consumption of Lighting (kWh) | Outdoor Temperature (°C) |
---|---|---|---|---|
Cluster Ⅰ: closed | 25 | C | 1190.13 | 20.79 |
MAX | 1329.90 | 24.95 | ||
MIN | 805.47 | 17.35 | ||
Cluster Ⅱ: slightly open | 17 | C | 1495.04 | 24.62 |
MAX | 1686.75 | 30.47 | ||
MIN | 1353.89 | 22.41 | ||
Cluster Ⅲ: severely open | 13 | C | 2066.55 | 28.84 |
MAX | 2280.90 | 32.05 | ||
MIN | 1787.04 | 25.38 |
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. |
© 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
Share and Cite
Zhao, T.; Zhang, C.; Ujeed, T.; Ma, L. Online Methodology for Separating the Power Consumption of Lighting Sockets and Air-Conditioning in Public Buildings Based on an Outdoor Temperature Partition Model and Historical Energy Consumption Data. Appl. Sci. 2021, 11, 1031. https://doi.org/10.3390/app11031031
Zhao T, Zhang C, Ujeed T, Ma L. Online Methodology for Separating the Power Consumption of Lighting Sockets and Air-Conditioning in Public Buildings Based on an Outdoor Temperature Partition Model and Historical Energy Consumption Data. Applied Sciences. 2021; 11(3):1031. https://doi.org/10.3390/app11031031
Chicago/Turabian StyleZhao, Tianyi, Chengyu Zhang, Terigele Ujeed, and Liangdong Ma. 2021. "Online Methodology for Separating the Power Consumption of Lighting Sockets and Air-Conditioning in Public Buildings Based on an Outdoor Temperature Partition Model and Historical Energy Consumption Data" Applied Sciences 11, no. 3: 1031. https://doi.org/10.3390/app11031031
APA StyleZhao, T., Zhang, C., Ujeed, T., & Ma, L. (2021). Online Methodology for Separating the Power Consumption of Lighting Sockets and Air-Conditioning in Public Buildings Based on an Outdoor Temperature Partition Model and Historical Energy Consumption Data. Applied Sciences, 11(3), 1031. https://doi.org/10.3390/app11031031