Please wait a minute...
浙江大学学报(工学版)
Journal of ZheJiang University (Engineering Science)  2020, Vol. 54 Issue (9): 1677-1689    DOI: 10.3785/j.issn.1008-973X.2020.09.003
    
Analysis of air-conditioning usage and energy consumption in campus teaching buildings with data mining
Xin-yue LI1(),Shu-qin CHEN1,*(),Hong-liang LI2,3,Yun-xiao LOU2,Jia-he LI4
1. College of Civil Engineering and Architecture, Zhejiang University, Hangzhou 310058, China
2. Zhejiang Excenergy Energy-saving Technology Co. Ltd, Hangzhou 310052, China
3. College of Control Science and Engineering, Zhejiang University, Hangzhou 310027, China
4. Zhejiang Bluetron Industry Internet Information Technology Co. Ltd, 310053, China
Download: HTML     PDF(1289KB) HTML
Export: BibTeX | EndNote (RIS)      

Abstract  

This study was based on the real-time operation data of the air-conditioning (AC) system which were collected by the energy consumption monitoring platform in a university located in Zhejiang province from November 2016 to February 2019. With the clustering analysis, six typical AC usage patterns and four energy consumption patterns were proposed for a whole year. Two supervised machine learning methods, namely the decision tree and random forest, were used to decouple the relation of AC usage and its energy consumption, and to figure out the different energy consumption levels under different AC usage conditions. The cross-validation method was used to compare the accuracy of various machine learning algorithms. Results show that the usage hour directly influences the energy consumption. The area scale of classrooms and AC use intensity have the significant effect on energy consumption in cooling scenario. The results of this study are beneficial to energy-saving management and the simulation of energy consumption for the teaching buildings in colleges and universities.

Key words： teaching buildings      data mining      air-conditioner usage      air-conditioning energy consumption      decoupling influences of energy consumption     
Received: 11 September 2019      Published: 22 September 2020
CLC:  TU 201.5  
Corresponding Authors: Shu-qin CHEN     E-mail: lixinyue@zju.edu.cn;hn_csq@126.com
Service
E-mail this article
Add to my bookshelf
Add to citation manager
E-mail Alert
RSS
Articles by authors
Xin-yue LI
Shu-qin CHEN
Hong-liang LI
Yun-xiao LOU
Jia-he LI
Cite this article:

Xin-yue LI,Shu-qin CHEN,Hong-liang LI,Yun-xiao LOU,Jia-he LI. Analysis of air-conditioning usage and energy consumption in campus teaching buildings with data mining. Journal of ZheJiang University (Engineering Science), 2020, 54(9): 1677-1689.

URL:

http://www.zjujournals.com/eng/10.3785/j.issn.1008-973X.2020.09.003     OR     http://www.zjujournals.com/eng/Y2020/V54/I9/1677


应用数据挖掘的高校教学建筑空调使用及其能耗分析

以浙江省某高校为研究对象，根据其节能监管平台在2016年11月—2019年2月的空调实时运行数据，利用聚类方法，全年共得到6种典型空调使用模式以及4种空调能耗模式. 利用基于监督学习的决策树、随机森林算法，对空调使用与能耗的关系进行解耦，明确不同空调使用情况导致的不同能耗水平，并使用交叉验证的方法比较多种机器学习算法的精度. 分析结果表明：空调使用时长均直接影响着日空调能耗，且在制冷工况下教室规模和空调使用强度也对能耗有着明显的影响. 研究结果可为高校教学建筑的节能管理及其能耗模拟提供支持.


关键词： 教学建筑,  数据挖掘,  空调使用行为,  空调能耗,  能耗影响解耦 
N/人 Rroom/% Smax Smin NAC Ruse
20 17.1 40 25 1 0.112
40 21.3 50 30 1 0.093
60 18.7 64 70 2 0.125
90 36.1 85 100 3 0.115
>90 2.9 460 115 4~16 0.132
Tab.1 Classroom information statistical table of certain university in Zhejiang
Fig.1 Air-conditioning service time distribution of different scale classrooms
Fig.2 Monthly mean outdoor temperature and average values of heating/cooling ratio from year of 2016 to 2019
季节 月份 空调工况 校历安排
夏初、夏末 6、9 制冷 教学
盛夏 7、8 制冷 暑假
过渡季 4、5、10、11 制冷、制热 教学
冬初、冬末 12、3 制热 教学
严冬 1、2 制热 寒假
Tab.2 Division of air-conditioning usage period in whole year
时间 日程
8:05 — 12:10 上午课程
12:10 — 13:10 午饭及休息
13:10 — 16:40 下午课程
16:40 — 18:00 晚饭及休息
18:00 — 20:25 晚上课程
Tab.3 Schedules of university in Zhejiang
Fig.3 Framework of analysis on air conditioning usage and energy consumption
类1 ··· 类i ··· 类m
类1 n1，1 … n1，i … n1，m
? ? ? ?
类i ni，1 … ni，i … ni，m
? ? ? ?
类m nm，1 … nm，i … nm，m
Tab.4 Schematic of confusion matrix to evaluate energy consumption decoupling effect
Fig.4 Hourly usage rate of each typical air-conditioning usage pattern
聚类 典型模式 P Tuse/h
P1 间断使用 0.298 2.10
P2 全时段使用 0.159 12.06
P3 白天使用 0.247 7.96
P4 下午使用 0.138 3.68
P5 上午使用 0.092 4.87
P6 后半天使用 0.066 7.56
Tab.5 Characteristic of air-conditioning usage patterns
季节 空调工况 下午使用 全时段使用 白天使用 上午使用 后半天使用 间断使用
盛夏 制冷 ? 0.17 0.39 ? ? 0.44
夏初、夏末 制冷 0.30 0.10 0.14 0.15 0.10 0.21
过渡季 制冷 0.21 0.14 0.17 0.15 0.10 0.24
过渡季 制热 ? 0.29 0.17 ? 0.14 0.40
严冬 制热 ? 0.32 0.28 ? ? 0.40
冬初、冬末 制热 0.16 0.11 0.25 0.16 0.11 0.22
Tab.6 Proportions of typical air-conditioning usage patterns in each season
Fig.5 Distribution of air-conditioning energy consumption in different classrooms with different capacities
Fig.6 Distributions of energy use intensity in the classrooms with different capacities
空调工况 能耗模式 P Tuse/h E/（kW·h） S/（kW·h） IAU
制冷 低能耗 0.317 2.30 4.03 0.67 0.59
中等能耗 0.264 4.90 16.65 1.92 0.66
长时间、中等能耗 0.229 10.25 26.58 1.37 0.66
长时间、高能耗 0.189 9.16 66.30 4.22 0.76
制热 低能耗 0.293 2.48 6.77 1.09 0.51
中等能耗 0.249 4.30 26.50 3.17 0.61
长时间、中等能耗 0.203 9.25 28.46 1.71 0.48
长时间、高能耗 0.255 9.08 74.19 4.54 0.70
Tab.7 Eigenvalues of typical energy consumption patterns under cooling and heating scenarios
Fig.7 Distributions of mean outdoor temperature and air-conditioning set temperature in each month
Fig.8 Distribution of energy patterns under each usage pattern in cooling scenario
Fig.9 Distribution of energy pattern under each usage patterns in heating scenario
Fig.10 Decision tree model of energy consumption decoupling in cooling scenario
Fig.11 Decision tree model of energy consumption decoupling in heating scenario
HE_LT ME_LT ME LE
HE_LT 202 93 0 0
ME_LT 222 566 0 0
ME 122 2 581 110
LE 9 0 204 806
Tab.8 Confusion matrix of decision tree model in cooling scenario
HE_LT ME_LT ME LE
HE_LT 318 97 12 1
ME_LT 139 253 29 3
ME 51 26 375 230
LE 0 0 52 376
Tab.9 Confusion matrix of the decision tree model in heating scenario
HE_LT ME_LT ME LE
HE_LT 262 72 41 1
ME_LT 224 589 0 0
ME 67 0 577 129
LE 2 0 167 786
Tab.10 Confusion matrix of random forest model in cooling scenario
HE_LT ME_LT ME LE
HE_LT 343 99 26 4
ME_LT 105 273 14 3
ME 59 13 358 134
LE 1 0 88 442
Tab.11 Confusion matrix of random forest model in heating scenario
工况 能耗模式 Rpre Rre
制热 LE 0.832 0.758
ME 0.635 0.737
ME_LT 0.691 0.709
HE_LT 0.727 0.675
制冷 LE 0.823 0.858
ME 0.746 0.735
ME_LT 0.724 0.891
HE_LT 0.697 0.672
Tab.12 Statistical index of random forest model in cooling and heating scenarios
工况 Tuse Puse IAU Scls Ss
制冷 2040.53 1028.93 382.22 302.25 172.12
制热 1144.21 576.24 286.33 132.28 152.33
Tab.13 Gini coefficient of each variable in cooling and heating scenarios
[1]   谭洪卫, 徐钰琳, 胡承益, 等 全球气候变化应对与我国高校校园建筑节能监管[J]. 建筑热能通风空调, 2010, 29 (1): 36- 40
TAN Hong-wei, XU Yu-lin, HU Cheng-yi, et al Research on building campus energy management[J]. Building Energy and Environment, 2010, 29 (1): 36- 40
doi: 10.3969/j.issn.1003-0344.2010.01.009
[2]   赵美, 于航, 石磊, 等 上海某高校学院楼基于能耗审计结果的节能潜力分析[J]. 建筑节能, 2017, 45 (4): 100- 104
ZHAO Mei, YU Hang, SHI Lei, et al Energy-saving potential analysis of a college building in shanghai based on the result of energy consumption audit[J]. Building Energy Efficiency, 2017, 45 (4): 100- 104
doi: 10.3969/j.issn.1673-7237.2017.04.020
[3]   高彪, 谭洪卫, 宋亚超 高校校园建筑用能现状及存在问题分析: 以长三角地区某综合型大学为例[J]. 建筑节能, 2011, 39 (2): 41- 44
GAO Biao, TAN Hong-wei, SONG Ya-chao Campus building energy consumption: taking one comprehensive university as example[J]. Building Energy Efficiency, 2011, 39 (2): 41- 44
doi: 10.3969/j.issn.1673-7237.2011.02.011
[4]   D ’OCA S, HONG T Occupancy schedules learning process through a data mining framework[J]. Energy and Buildings, 2015, 88 (1): 395- 408
[5]   LIANG X, HONG T, SHEN G Occupancy data analytics and prediction: a case study[J]. Building and Environment, 2016, 102 (6): 179- 192
[6]   D’OCA S, HONG T A data-mining approach to discover patterns of window opening and closing behavior in offices[J]. Building and Environment, 2014, 82 (12): 726- 739
[7]   胡天乐, 曹彬, 朱颖心, 等 北京高校自习教室使用率对供暖季室内热环境的影响[J]. 暖通空调, 2015, 45 (8): 98- 101
HU Tian-le, CAO Bin, ZHU Yin-xin, et al Influence of usage rate on indoor thermal environment of university ’s self-learning classrooms in Beijing in heating season[J]. Heating Ventilating and Air Conditioning, 2015, 45 (8): 98- 101
[8]   SONG K, KIM S, PARK M, et al Energy efficiency-based course timetabling for university buildings[J]. Energy, 2017, 139 (11): 394- 405
[9]   郝乐, 郭广翠 高校节能监管平台建设现状及发展展望[J]. 建设科技, 2017, (14): 63- 65
HAO Le, GUO Guang-cui State and prospect of energy-saving monitoring platform in high school[J]. Construction Science and Technology, 2017, (14): 63- 65
[10]   ZHAO Y, XIAO F, WANG S An intelligent chiller fault detection and diagnosis methodology using Bayesian belief network[J]. Energy and Buildings, 2013, 57 (2): 278- 288
[11]   LI G, CHEN H, HU Y, et al An improved decision tree-based fault diagnosis method for practical variable refrigerant flow system using virtual sensor-based fault indicators[J]. Applied Thermal Engineering, 2017, 129 (1): 1292- 1303
[12]   XIAO F, FAN C Data mining in building automation system for improving building operational performance[J]. Energy and Buildings, 2014, 75 (6): 109- 118
[13]   肖赋, 范成, 王盛卫 基于数据挖掘技术的建筑系统性能诊断和优化[J]. 化工学报, 2014, 44 (6): 1188- 1198
XIAO Fu, FAN Cheng, WANG Sheng-wei Building system performance diagnosis and optimization based on data mining techniques[J]. Journal of Chemical Industry and Engineering, 2014, 44 (6): 1188- 1198
[14]   REN X, YAN D, HONG T Data mining of space heating system performance in affordable housing[J]. Building and Environment, 2015, 89 (7): 1- 13
[15]   谈雪, 郭强, 许健, 等 夏热冬冷地区某高校典型建筑用能特征与用能行为影响分析[J]. 南京理工大学学报, 2019, 43 (1): 101- 107
TAN Xue, GUO Qiang, XU Jian, et al Characteristics of energy use of typical buildings and its behavioral influencing analysis in one university campus in hot summer and cold winter[J]. Journal of Nanjing University of Science and Technology, 2019, 43 (1): 101- 107
[16]   张喜勇. 区域建筑空调负荷预测和特征分析方法研究[D]. 杭州: 浙江大学, 2018
ZHANG Xi-yong. Prediction and analysis method of air-conditioning load in regional buildings[D]. Hangzhou: Zhejiang University, 2018.
[17]   LI X, CHEN S, LI J, et al. Air-conditioner usage patterns in teaching buildings of universities by data mining approach [C] // Proceedings of the 11th International Symposium on Heating, Ventilation and Air Conditioning (ISHVAC 2019). Harbin: ISHVAC, 2019: 649-659
[18]   李寅雷, 李鸿亮, 蔡强新, 等 校园多联机空调的计费监控系统集成应用[J]. 浙江建筑, 2016, 33 (9): 48- 50
LI Ying-lei, LI Hong-liang, CAI Qiang-xin, et al Integrated application of the charge monitoring system for the multi connected air conditioner in campus[J]. Zhejiang Construction, 2016, 33 (9): 48- 50
doi: 10.3969/j.issn.1008-3707.2016.09.012
[19]   GU J, XU P, PANG Z, et al Extracting typical occupancy data of different buildings from mobile positioning data[J]. Energy and Buildings, 2018, 180 (12): 135- 145
[20]   陈义波. 数据驱动的典型公共建筑多维度能耗预测模型研究[D]. 上海: 同济大学, 2018.
CHEN Yi-bo. Data-driven prediction models of multi- dimensional energy consumed in typical public buildings[D]. Shanghai: Tongji University, 2018.
[21]   MEESRIKAMOLKUL W, NIENNATTRAKUL V, RATANAMAHATANA C A. Shape-based clustering for time series data [C] // Pacific-Asia Conference on Knowledge Discovery and Data Mining. Berlin, Heidelberg: PAKDD, 2012: 530-541.
[22]   HAN J. Data Mining: Concepts and Techniques [M]. San Francisco: Morgan Kaufmann Publishers Inc. 2005.
[23]   YU Z, HAGHIGHAT F, FUNG B C M, et al A decision tree method for building energy demand modeling[J]. Energy and Buildings, 2010, 42 (10): 1637- 1646
doi: 10.1016/j.enbuild.2010.04.006
[1] Jin LIANG,Kun LUO,Qiang WANG,Xu-chao YANG,Jian-ren FAN,Jun-xi ZHANG. Effect of high-albedo roofs on urban heat island and air-conditioning energy consumption[J]. Journal of ZheJiang University (Engineering Science), 2020, 54(10): 1993-2000.
[2] WANG Cheng-long, LI Cheng, FENG Yi-ping, RONG Gang. Dispatching rule extraction method for job shop scheduling problem[J]. Journal of ZheJiang University (Engineering Science), 2015, 49(3): 421-429.
[3] WANG Xiang-bing,TONG Shui-guang,ZHONG Wei,ZHANG Jian. Study on  scheme design technique for hydraulic excavator's structure performance based on extension reuse[J]. Journal of ZheJiang University (Engineering Science), 2013, 47(11): 1992-2002.
[4] MA Jin, Li-Feng, LI Jian-Hua. Perturbation method for distributed privacy-preserving data mining[J]. Journal of ZheJiang University (Engineering Science), 2010, 44(2): 276-282.
[5] DAN Fu-Dong, CA Ming, TUN Hong-Sen, DONG Jin-Xiang, FU Gao. An incremental algorithm for mining frequent closed patterns[J]. Journal of ZheJiang University (Engineering Science), 2009, 43(8): 1389-1395.
[6] WANG Guang-Ceng, CAO Yi-Jia, CU Jun, et al. Design and implementation of general report integrated  management system for power company[J]. Journal of ZheJiang University (Engineering Science), 2009, 43(11): 2062-2066.