多模型融合的民机环控系统空调组件故障诊断模型开发及验证  

Development and Validation of Multi-model Fusion Fault Diagnosis Model for Air-conditioning Packs of Civil Aircraft Environmental Control System

作　　者：胡鸣鹤 李昱翰 张弛 胡海涛[1] 孙浩然 吴成云 舒悦 HU Minghe;LI Yuhan;ZHANG Chi;HU Haitao;SUN Haoran;WU Chengyun;SHU Yue(Institute of Refrigeration and Cryogenics,Shanghai Jiao Tong University,Shanghai 200240,China;Shanghai Aircraft Design and Research Institute,Commercial Aircraft Corporation of China,Co.,Ltd.,Shanghai 201210,China;State Key Laboratory of Compressor Technology(Anhui Laboratory of Compressor Technology),Hefei 230031,Anhui,China)

机构地区：[1]上海交通大学制冷与低温工程研究所,上海200240 [2]中国商用飞机有限责任公司上海飞机设计研究院,上海201210 [3]压缩机技术国家重点实验室(压缩机技术安徽省实验室),安徽合肥230031

出　　处：《制冷技术》2024年第3期29-35,共7页Chinese Journal of Refrigeration Technology

基　　金：上海市科委科技创新行动计划(No.19142203000);压缩机技术国家重点实验室(No.SKLYSJ201904);上海市启明星计划杨帆专项(No.22YF1459700)。

摘　　要：针对民机环控系统空调组件故障数据难获取、故障数据不平衡的问题,本文采用动态仿真的方法,研究了民机环控系统空调组件的故障发生情况,通过人工注入故障获取了故障数据集。利用滑动平均和经验模态分解的方法有效抑制了空调组件原始时序信号中的白噪声,降低了将正常信号误判为异常的风险。开发了融合孤立森林、谱残差和自编码器的故障诊断模型,分析了不同算法组合的融合算法精度。结果表明,开发的融合模型可以有效诊断动态仿真数据集中的空调组件故障,诊断精度达到100%;对于机械故障公开数据集,训练的模型结合开发的信号处理方法可以实现91.12%的诊断准确率,召回率达到98.66%,验证了解决方案的有效性和泛用性。Aiming at the difficulty to obtain fault data and the unbalanced fault data composition of the air-conditioning packs in the civil aircraft environmental control system,the method of dynamic simulation to study the occurrence of faults,and obtains the fault data set by manually injecting fault conditions is adopted in this paper.The white noise in the original timing signals of the air-conditioning packs is effectively suppressed using sliding average and empirical modal decomposition,which reduces the risk of misjudging normal signals as abnormal.A fault diagnosis model integrating isolated forest,spectral residual,and autoencoder is developed,and the accuracy of the combined model is analyzed for different combinations of algorithms.The results show that,the developed model can effectively diagnose air-conditioning packs faults in the dynamic simulation dataset with a diagnostic precision of 100%;for the open dataset of mechanical faults,the trained model combined with the developed signal processing method can achieve a diagnostic accuracy of 91.12%,and the recall rate reaches 98.66%,which verifies the effectiveness and ubiquity of the solution.

关 键 词：空调组件 去噪 模型 故障诊断 

分 类 号：TU834[建筑科学—供热、供燃气、通风及空调工程] TP306.3[自动化与计算机技术—计算机系统结构]