Objective The fault detection results of the selected parameters require effective fusion to obtain comprehensive system detection results to accurately monitor faults occurring during the operation of the air-conditioning and refrigeration system. Considering the nonlinear and non-Gaussian variation characteristics of temperature and pressure in the selected parameter variables, extracting the abnormal fluctuation characteristics of the air-conditioning and refrigeration system from individual variables becomes the primary problem addressed. Because single-variable monitoring data are susceptible to wild values, noise, and other disturbances, relying only on a single variable to detect faults in the air-conditioning system can produce inaccurate results. Therefore, integrating the fluctuation characteristics of multiple related variables of the air-conditioning and refrigeration system based on the abnormal fluctuation characteristics of each variable using appropriate methods to obtain the overall fault detection results of the system, becomes another problem addressed in this study. Therefore, this research proposes a multivariate fluctuation fusion-oriented fault detection method for air-conditioning and refrigeration systems based on the research concept of "single-variable first, and then overall". Methods This study proposed a fault detection method for air conditioning and refrigeration systems based on the fusion of multivariate fluctuation features. First, the abnormal fluctuation characteristics of a single variable were extracted through prediction and smoothing techniques to achieve the Gaussianization of non-Gaussian variables and obtain fault detection evidence that reflected the fluctuation characteristics of the variable. Second, the nonlinear correlation among multiple variables in the air conditioning system was captured using the distance correlation coefficient (dCor). Finally, the evidential reasoning rule with dependent evidence (ERr‒DE) method was utilized to fuse the fluctuation characteristic evidence of multiple related variables in the air conditioning system to obtain the overall fault detection results at the system level. The main innovations of the study included 1) the proposed fluctuation feature extraction method for non-Gaussian univariate variables based on prediction and smoothing techniques, which achieved the Gaussianization transformation of monitoring data for non-Gaussian variables in the air conditioning system, extracted the fluctuation features of univariate variables, and provided the basis for fault detection at the univariate level, and 2) the adoption of a correlation evidence inference method to effectively fuse fluctuation features of multiple relevant variables in the air conditioning system to obtain overall fault detection results at the system level, which provided a new solution for fault detection in air conditioning and refrigeration systems. Results and Discussions This study verified the effectiveness of the proposed method through experiments and analyzed its advantages in detecting system faults at different levels. It also provided a basis for optimizing the operation and maintenance management of air conditioning and refrigeration systems. In this study, fuzzy inference theory, deep learning, and confidence rule base methods were selected as comparative approaches for fault monitoring. Compared to fuzzy inference theory and neural network theory, the diagnostic accuracy improved by 15.75% and 13.75%, respectively. Compared to the confidence rule base method, the accuracy improved by 30.75%. Therefore, the comparison experiments indicated that the performance of the diagnostic model based on the ERr-DE method was superior to other methods. In addition, the method proposed in this study not only obtained fault detection results for individual system variables but also fused them to generate overall fault detection results for the system. This capability allowed system managers to comprehensively understand the fault status of the system at different levels. Specifically, the single-variable fault detection results were utilized to detect early signs of system abnormalities, providing alarms for front-line duty personnel. In addition, the fused overall fault detection results were utilized to assess the overall operational status of the system, which provided a basis for decision-making when optimizing system operation and maintenance strategies. Conclusions This study proposes a new fault detection method for air-conditioning and refrigeration systems. The method is based on prediction and smoothing techniques to extract the fluctuation characteristics of the system's single-variable time series. On this basis, the ERr‒DE model is employed to achieve effective fusion of evidence generated by the refrigeration system's multi-correlation variables and to obtain the overall fault detection results of the system. The proposed method is applied to a fault detection experiment of a certain type of air-conditioning refrigeration system to verify its effectiveness. The obtained results enable system operation and maintenance managers to understand the fault status of the system at different levels, providing a reliable decision-making basis for early abnormal alarms and the overall operation and maintenance of the system.
LaneK, Smalls‒ManteyL, HernándezD,et al.Extreme heat and COVID-19 in New York City:An evaluation of a large air conditioner distribution program to address compounded public health risks in summer 2020[J].Journal of Urban Health,2023,100(2):290-302. doi:10.1007/s11524-022-00704-9
[2]
DingYifan, XiWeimin, MaYahui,et al.Research on large-scale air-conditioning loads participating in peak shaving[J].Earth and Environmental Science,2021,831(1):012032. doi:10.1088/1755-1315/831/1/012032
[3]
LiGuolin.Research on air conditioner fault prediction based on deep learning[D].Chengdu:Southwest Petroleum University,2019.
[4]
李果林.基于深度学习的空调故障预测研究[D].成都:西南石油大学,2019.
[5]
MuKang, LiuJing, LuZhen,et al.Analysis of heat rejection from an actual large-scale air-conditioned office building by field measurements and numerical simulations[J].Energy and Buildings,2016,111:369-379. doi:10.1016/j.enbuild.2015.11.035
[6]
HuangYueheng.Large air conditioning unit equipment fault analysis and health management system technology research[D].Xi’an:Xidian University,2022.
[7]
黄悦恒.大型空调机组设备故障分析与健康管理系统技术研究[D].西安:西安电子科技大学,2022.
[8]
Babadi SoultanzadehM, OufM M, Nik-BakhtM,et al.Fault detection and diagnosis in light commercial buildings’ HVAC systems:A comprehensive framework,application,and performance evaluation[J].Energy and Buildings,2024,316:114341. doi:10.1016/j.enbuild.2024.114341
[9]
FanCheng, HeWeilin, LiuYichen,et al.A novel image-based transfer learning framework for cross-domain HVAC fault diagnosis:From multi-source data integration to knowledge sharing strategies[J].Energy and Buildings,2022,262:111995. doi:10.1016/j.enbuild.2022.111995
[10]
SulaimanN.Fault detection and diagnosis of air-conditioning system using long short-term memory recurrent neural network[J].Przegląd Elektrotechniczny,2023,1(9):115-119. doi:10.15199/48.2023.09.21
[11]
TudoroiuN, ZaheeruddinM, TudoroiuE R,et al.Fault detection and diagnosis(FDD) in heating ventilation air conditioning systems(HVAC) using an interactive multiple model augmented unscented Kalman filter(IMMAUKF)[C]//Proceedings of the 2008 Conference on Human System Interactions.Krakow:IEEE,2008:334-339. doi:10.1109/HSI.2008.4581459
[12]
LeeK P, WuB H, PengShilin.Deep-learning-based fault detection and diagnosis of air-handling units[J].Building and Environment,2019,157:24-33. doi:10.1016/j.buildenv.2019.04.029
[13]
YangSiyu, YangXuebin, LiXinhai,et al.On-line training strategy for valve stuck fault detection threshold of VAV air conditioning[J].Journal of Donghua University(Natural Science Edition),2023,49(4):123-129. doi:10.19886/j.cnki.dhdz.2022.0018
LiTingting.Research on general fault detection and diagnosis methods for heating,ventilation and air conditioning systems[D].Hangzhou:Zhejiang University,2022.
[16]
李婷婷.暖通空调系统通用性故障检测与诊断方法研究[D].杭州:浙江大学,2022.
[17]
ZhangChaobo.Research on knowledge-and-data-driven fault detection and diagnosis methods for heating, ventilation and air conditioning systems[D].Hangzhou:Zhejiang University,2022.
[18]
章超波.知识-数据驱动的暖通空调系统故障检测与诊断方法研究[D].杭州:浙江大学,2022.
[19]
WangShuting.Research on fault detection and diagnosis of all air air conditioning systems using bayesian network-based diagnostic units[D].Hangzhou:Zhejiang University,2021.
[20]
王姝婷.基于贝叶斯网络单元的全空气空调系统故障检测与诊断方法研究[D].杭州:浙江大学,2021.
[21]
ZhangHui, YangXuebin, FangXing,et al.A fault detection and diagnosis technology for chilled water valve based on Bayesian network[J].Journal of Donghua University(Natural Science Edition),2020,46(1):141-147.
WangYuan.Research on fault detection and diagnosis strategy of vav air-conditioning system:A case study of five AHU faults[D].Xi'an:Chang'an University,2020.
ChengDaheng.Experimental and theoretical research on integrated fault detection and diagnosis model for two faults in VRF System based on 1-D CNN[D].Wuhan:Huazhong University of Science and Technology,2020.
SunXueteng.Research on air-condition fault detection algorithm based on deep learning[D].Beijing:China Jiliang University,2022.
[30]
孙学腾.基于深度学习的空调故障检测算法研究[D].北京:中国计量大学,2022.
[31]
ShanBiao, DuJun, ShangLiangliang.Fault detection and diagnosis of sensors in air-conditioning system based on improved PCA method[J].Control Engineering of China, 2020,27(4):765-770.
ZhangShuangshuang, ChenHuanxin, ZhangHongtao,et al.Sensor fault detection and diagnosis of air-conditioning system based on improved principal component analysis method[J].Journal of Refrigeration,2020,41(1):147-153.
GuoPengwei, MengWeina, BaoYi.Knowledge graph-guided data-driven design of ultra-high-performance concrete (UHPC) with interpretability and physicochemical reaction discovery capability[J].Construction and Building Materials,2024,430:136502. doi:10.1016/j.conbuildmat.2024.136502
[36]
LiXiayang, LiuChang, YangXiaoyfeng,et al.A fault diagnosis strategy for track circuits based on data-knowledge collaboratively driven[J].Journal of Railway Science and Engineering,2024,21(12):5276-5288.
ProvostF, FawcettT.Data science and its relationship to big data and data-driven decision making[J].Big Data,2013,1(1):51-59. doi:10.1089/big.2013.1508
[39]
XuDandan, ZhangDi.Diagnosis and inference of spindle faults in CNC machine tools based on data-driven and ontology modeling[J].Machine Tool & Hydraulics,2024,52(12):244-252. doi:10.3969/j.issn.1001-3881.2024.12.035
WengXu, XuXiaobin, FengJing,et al.A correlation analysis-based multivariate alarm method with maximum likelihood evidential reasoning[J].IEEE Transactions on Automation Science and Engineering,2024,21(4):4974-4986. doi:10.1109/tase.2023.3305524
[45]
ZhangPeng, ZhouZhijie, TangShuaiwen,et al.On the evidential reasoning rule for dependent evidence combination[J].Chinese Journal of Aeronautics,2023,36(5):306-327. doi:10.1016/j.cja.2023.03.037
[46]
曾青中,邓景山.车辆空调与制冷装置[M].成都:西南交通大学出版社,2008.
[47]
ZhangGuangye.Study on failure and fault diagnosis of urban rail vehicle air conditioning system[D].Beijing:Beijing University of Civil Engineering and Architecture,2013.
[48]
张广也.城轨车辆空调制冷系统的失效及故障诊断研究[D].北京:北京建筑大学,2013.
[49]
WuYanhua.Research on the AFDR-based multivariate EWMA chart[D].Ji'nan:Shandong University of Finance and Economics,2017.
LiJia, DiShimin, ShenYanyan,et al.FluxEV:A fast and effective unsupervised framework for time-series anomaly detection[C]//Proceedings of the 14th ACM International Conference on Web Search and Data Mining.New York:ACM,2021:824-832. doi:10.1145/3437963.3441823
[52]
YangJianbo.Rule and utility based evidential reasoning approach for multiattribute decision analysis under uncertainties[J].European Journal of Operational Research,2001,131(1):31-61. doi:10.1016/s0377-2217(99)00441-5
[53]
TangShuaiwen, ZhouZhijie, HuChanghua,et al.A new evidential reasoning rule-based safety assessment method with sensor reliability for complex systems[J].IEEE Transactions on Cybernetics,2022,52(5):4027-4038. doi:10.1109/tcyb.2020.3015664