故障诊断与寿命预测是解决PEMFC系统寿命短、成本高的有效方法。综述了PEM系统的退化机制,包括质子交换膜、催化层、扩散层等各个组件的退化机理。从本征因素和外部因素两方面总结了影响PEM系统老化的影响因素,特别是在运行管理过程中操作不当会加速PEM系统老化。介绍了表征PEM老化程度的常用退化性能指标,退化性能指标的选择对寿命预测方法的研究有重要意义,介绍了不同指标的优缺点以及适用情况。从模型驱动方法、数据驱动方法和混合驱动方法3方面介绍了常见的PEM系统寿命预测方法。模型驱动和数据驱动方法各有优缺点,混合方法可结合二者优点并提高预测的精度和准确性。最后展望了预测研究的未来发展方向。

Remaining useful life prediction is an effective solution to solve the problem of short life and high cost of PEMFC system. In thispaper, the degradation mechanism of PEM system, including proton exchange membrane, catalytic layer, diffusion layer and other components, was reviewed. The factors influencing the aging of PEM system were summarized from intrinsic factors and external factors, especially improper operation in operation and management would accelerate the aging of PEM system. The commonly used degradation performance indexes to characterize the aging degree of PEM were summarized. The selection of degradation performance indexes is of greatsignificance to the research of life prediction. The advantages, disadvantages and applicability of different indexes were introduced. Common faults and life prediction were introduced from three aspects: model-driven, data-driven and hybrid-driven. Model-driven and datadriven have their own advantages and disadvantages. Hybrid can combine their advantages and improve the precision and accuracy of prediction. Finally, the future direction of prediction research was prospected.

0 引言

1 PEMFC衰退机制

   1.1 PEMFC衰退机制

   1.2 催化层衰退机制

   1.3 扩散层衰退机制

2 退化影响因素

   2.1 不同工况的影响

   2.2 运行条件管理的影响

3 退化性能指标

4 寿命预测方法

   4.1 基于物理模型的预测方法

   4.2 数据驱动方法

   4.3 混合驱动

5 结语及展望