基于关键特征选择的冷凝器脏堵故障诊断OA

To address the rise in air conditioning energy consumption induced by condenser fouling and blockage faults of electric vehicles,a fault diagnosis method integrating heat transfer mechanisms and back propagation neural networks is proposed in this paper.Through the analysis and simulation verification of the heat transfer mechanism of the condenser,the heat transfer temperature difference is identified as a sensitive feature parameter to characterize the fault,and the importance of each feature to the heat transfer temperature difference is analyzed based on the knowledge of thermal engineering,and the key feature parameters affecting the heat transfer temperature difference are selected,the back-propagation neural network model is established to predict the heat transfer temperature difference using the normal sample data,and fault diagnosis is carried out through the residual analysis.The results show that the method not only reduces the number of sensor requirements and data storage space,but also trains a model with high prediction accuracy and a coefficient of determination of 0.984,which provides an effective support for the early warning of the condenser dirty plugging fault.