机载电子设备的故障诊断和趋势预测

Fault Diagnosis and Prediction for Airborne Electronic Equipment

提高机载电子设备的可靠性,已成为保证飞机性能的当务之急。本文首先指出机载电子设备失效的主要原因是耗损性故障,接着分析了其常见的故障预测方法,最后给出了采用时间序列自回归(AR)模型和基于支持向量回归的时间序列预测(TS-SVR)模型对某机载雷达故障预测的实例。研究发现,AR模型仅适用于短期预测,而TS-SVR方法推广能力强、具有较强的鲁棒性和容错性,对较长区间预测仍具有较好的效果。因此,将TS-SVR方法应用于机载电子设备的故障预测,可以取得良好的结果。

It is necessary to study the ways to improve the reliability of airborne electronic equipment. In this paper, the failure law of airborne electronic equipment was introduced firstly. Then common failure prediction methods were summarized and analyzed. Finally, an example of predicting the airborne radar failure using the Auto-Regressive （AR） model and Support Vector Regression model based on Times Series （TS-SVR） was presented. According to numerical simulation, the AR forecasting method is only suitable to short-period prediction while the TS-SVR method possesses good robustness and fault-tolerant performance and applies to long-period prediction. Therefore, it is concluded that the TS-SRV method is feasible for fault diagnosis and prediction of airborne electronic equipment.