Health assessment method for civil aircraft air conditioning system integrating fault logic and Bayesian time-varying evaluation network

To effectively achieve health assessment of civil aircraft air conditioning system under limited sample conditions and address the challenge of evaluating state in the time dimension of air conditioning system, an assessment method that integrates fault logic with a Bayesian time-varying evaluation network is proposed. Firstly, based on the system operating principles, a three-level evaluation framework of "monitoring indicators—components—system" is constructed, and a fault logic diagram is employed to clarify the relationships between monitoring indicators and component states. Secondly, a time-varying comprehensively weighted fuzzy comprehensive evaluation method is established by introducing a temporal factor and information entropy independence weight coefficient coupling weighting strategy, enabling quantify health status of components in the time dimension. Thirdly, a game-equilibrium-driven Bayesian network is incorporated to map component states to the system state, forming the Bayesian time-varying coupling weighted fuzzy comprehensive evaluation model (BTCW-FCEM) for dynamic assessment of system-level health. Finally, case studies on single component and composite component faults are conducted using operational monitoring data from a domestic civil aircraft air conditioning system, validating the high accuracy and efficiency of model in dynamic health assessment. The results show that the BTCW-FCEM outperforms multiple comparative models in terms of accuracy, precision, recall, and F1 score in various fault scenarios, providing reliable technical support for health monitoring and fault warning of civil aircraft air conditioning systems.