针对飞机空调系统健康状态评价中物理信息、任务过程和环境应力关联弱导致的参数表征不完整、特征提取不完善和健康评价不准确问题，本文提出一种融合任务剖面与主成分分析-马氏距离（PCA-MD）的智能健康评价与退化预警方法，为航空装备复杂系统的智能健康评价提供一套通用的解决方案。首先，分析空调系统的功能结构与典型故障，构建包含整机级和系统级的健康表征参数体系，捕捉系统间的耦合失效特点；其次，分析飞行任务对空调系统的影响，设计基于任务剖面的健康参数谱，充分表达不同任务阶段健康参数的差异；再次，构建基于欧几里得范数的健康基线模型，综合考虑维修定检与环境变化对健康基线的影响；然后，提出基于PCA-MD的健康评价与退化预警方法，通过引入PCA解决健康指标构建中协方差矩阵不可逆和基于健康参数谱矩阵维度不一致的问题，实现健康指标的高效构建、健康状态的准确评价和性能退化的预警识别；最后，本文针对1474个实际A320飞机空调系统的数据样本开展应用验证研究，比较结果验证了本文方法的有效性、实用性与优越性。
Aiming at the problems of incomplete parameter characterization, imperfect feature extraction and inaccurate health evaluation caused by the weak correlation of physical information, mission process and environmental stress in the health state evaluation of aircraft air-conditioning system, this paper proposes an intelligent health evaluation and degradation warning method fusing mission profile and Principal Component Analysis- Mahalanobis Distance (PCA-MD) to provide a set of general solutions for intelligent health evaluation of complex systems of aviation equipment. First, this paper analyses the functional structure and typical failures of the air conditioning system, constructs a health characterization parameter system that includes the total level and the system level, and captures the coupled failure characteristics of the systems. Second, the impact of the flight mission on the air conditioning system is analyzed to design a health parameter spectrum based on the mission profile, which fully expresses the differences of the health parameters in the different mission phases. Third, a health baseline model based on Euclidean para-digm is constructed, which can comprehensively consider the influence of maintenance and inspection, as well as the environ-mental changes on the health baseline. Fourth, a health evaluation and degradation early warning method based on PCA-MD is proposed, which resolves the problem of the irreversibility of the covariance matrix in the construction of the health indexes and the inconsistency of the dimensions of the matrix based on the spectrum of the health parameters through the introduction of PCA, so as to achieve the efficient construction of health indexes, accurate evaluation of health state and early warning identifi-cation of performance degradation. Finally, this paper carries out an application validation study on the 1,474 practical data samples of the air conditioning system of a A320 airplane. The comparative results have validated the feasibility, practicability, and superiority of the proposed method.
[1] 钱长林.基于改进StemGNN的飞机空调系统故障分析研究 [D], 2022.
[2]石旭东, 蒋贵嘉, 张宇.基于联合仿真的飞机空调系统故障影响[J].航空学报, 2020, 41(08):295-303
[3]陈妍宇.飞机环控系统故障与健康状况预测方法研究 [D], 2019.
[4]李超役.民用飞机空调系统健康评估与故障诊断方法研究 [D], 2018.
[5]孙见忠, 解志峰, 闫洪胜.民机预测维修模式在空调系统的应用[J].南京航空航天大学学报, 2021, 53(06):952-64
[6]曹明, 黄金泉, 周健.民用航空发动机故障诊断与健康管理现状、挑战与机遇Ⅰ：气路、机械和系统故障诊断与预测[J].航空学报, 2022, 43(09):9-41
[7]曹明, 王鹏, 左洪福.民用航空发动机故障诊断与健康管理现状、挑战与机遇Ⅱ：地面综合诊断、寿命管理和智能维护维修决策[J].航空学报, 2022, 43(09):42-81
[8]孙见忠, 左洪福, 闫洪胜.民用飞机预测维修技术研究进展[J].航空科学技术, 2024, 35(07):14-31
[9]BILLINTON R, ALLAN R N.Reliability evaluation of engineering systems [M]. Springer, 1992.
[10]WANG C, ZHANG T, LUO F, et al.Fault incidence matrix based reliability evaluation method for complex distribution system[J].IEEE Transactions on Power Systems, 2018, 33(6):6736-45
[11]NUMAN M, BAIG M F, YOUSIF M.Reliability evalu-ation of energy storage systems combined with other grid flexibility options: A review [J].[J].Journal of Energy Storage, 2023, 63(107022):-
[12]PANDIAN G P, DIGANTA D, CHUAN L, et al.A cri-tique of reliability prediction techniques for avionics applications[J].Chinese Journal of Aeronautics, 2018, 31(1):10-20
[13]XU Y, KOHTZ S, BOAKYE J, et al.Physics-informed machine learning for reliability and systems safety ap-plications: State of the art and challenges [J].[J].Reliabil-ity Engineering & System Safety, 2023, 230(108900):-
[14]KHAN S, YAIRI T.A review on the application of deep learning in system health management [J].[J].Mechanical Systems and Signal Processing, 2018, 107:241-265
[15]陈志强, 陈旭东.深度学习在设备故障预测与健康管理中的应用[J].仪器仪表学报, 2019, 40(09):206-26
[16]CHE C, WANG H, FU Q, et al.Combining multiple deep learning algorithms for prognostic and health management of aircraft [J].[J].Aerospace Science and Technology, 2019, 94(105423):-
[17]房红征, 年夫强, 罗凯.基于机器学习建模的航天器健康管理平台研究[J].计算机测量与控制, 2022, 30(12):112-8
[18]REZAEIANJOUYBARI B, SHANG Y.Deep learning for prognostics and health management: State of the art, challenges, and opportunities [J].[J].Measurement, 2020, 163(107929):-
[19]ZHAO J, MA C, SHE Z.Research on Health Monitor-ing and Prediction Technology for Civil Aircraft Envi-ronmental Control Systems: A Review; proceedings of the ASME International Mechanical Engineering Con-gress and Exposition, F, 2023 [C]. American Society of Mechanical Engineers.
[20]曹曦丹.深度学习在B737引气系统健康管理中的应用研究 [D], 2022.
[21]王冉.基于QAR的航空发动机性能发展预测研究 [D], 2020.
[22]王奕首, 余映红, 卿新林.基于和的航空发动机排气温度基线模型[J].航空发动机, 2020, 46(01):54-60
[23]曾康, 赖凤琴, 周利敏等.民机整体驱动发电机典型故障分析与预防性维修策略优化 [J].[J].航空维修与工程, 2024, (07):79-83
[24]刘博.民航客机飞控系统健康管理关键技术研究 [D], 2017.
[25]蒋银, 李军, 汪志民.多功能安全分析平台在A320液压系统健康状态感知领域的应用研究 [J].[J].航空维修与工程, 2023, (03):56-8-
[26]方正汉.基于多特征的航电空调系统寿命预测研究 [D], 2020.
[27]陶言和, 郭勤涛, 周瑾.观测不确定性下变分贝叶斯高效模型修正[J].航空学报, 2024, 45(19):193-207
[28]邵怡韦, 陈嘉宇, 林翠颖.小训练样本下齿轮箱故障诊断：一种基于改进深度森林的方法[J].航空学报, 2022, 43(08):118-32
CJA