基于模型的系统级故障诊断与预测推理技术

A model based system level fault diagnosis and prognosis technology

针对装备状态监控和故障预测推理的需求,提出不同的诊断推理模型。首先,提出了基于二值相关性模型的快速推理算法,采用Vx Works平台对方法进行了验证评估。然后,针对应用过程中存在的时延、故障耦合等现象,提出了动态诊断模型。在此基础上,提出了考虑故障预测的诊断和预测一体化诊断推理模型,对推理运算方法进行了阐述,开发了仿真验证软件。以某卫星电源系统为例进行了验证,结果显示模型具有故障诊断和预测能力,能实现故障诊断和预测的一体化推理。

In this paper, several diagnostic models are discussed according to different needs of equipment status monitoring and fault prognosis. Firstly, a method based on binary dependency matrix is proposed and tested on a system with Vx Works. Then, for the problem of test delay and coupled faults, a dynamic diagnostic model is developed. After that, an integrated diagnostic model and inference algorithm for fault prediction is proposed. A simulation software is developed based on the method. The application result on a satellite shows that the model has the ability to realize unified fault diagnosis and prognosis.