两种多故障诊断算法的性能比较研究

Research on performance comparison of two MFD algorithms

多信号模型结合了结构模型与功能依赖模型的优点,非常适合于描述复杂系统故障传播关系。基于多信号模型故障推理算法非常重要,其中多故障诊断算法又是其中的难点。通过对基于多信号模型多故障诊断的拉格朗日松弛算法和主启发式算法及其在阿波罗飞船发射前系统状态诊断的应用中的深入对比研究,仿真结果表明拉格朗日松弛算法和主启发式算法均能有效解决多故障诊断问题,但主启发式算法的效率更高,尤其是在含有大量故障源和测试的大型实时系统中。

Multi-signal model,which combines the advantages of structure model and function dependency model,is very suitable for describing the complex relationship between fault propagation.The fault diagnosis algorithm based on multi-signal model is very important,especially the multiple fault diagnosis algorithm.Through the thorough comparative study between Lagrange relaxation algorithm and Primal heuristic algorithm which are both based on multi-signal model,as well as their application in the diagnosis of Apollo.Then the simulation present that both Lagrange relaxation algorithm and Primal heuristic algorithm can solve the multiple faulty diagnosis problem,while Primal heuristic algorithm has a high efficiency,especially in real-time diagnosis of large system with a large number of failure sourses and tests.