基于多信号模型航天器多故障诊断技术研究

最优多故障诊断问题是一个NP-COMPLETE问题。本文针对卫星电源这类复杂航天器系统，建立了多信号模型，该模型采用有向图描述元件与测试间的因果依赖关系。考虑系统元件发生故障的先验概率以及可用的测试工具，给出了一种基于拉格朗日松弛和子梯度优化算法的近最优多故障诊断算法。该建模方法与诊断方法在某卫星电源系统测试中得到了验证，结果表明该方法建模简单，诊断速度快，适用于航天器这种复杂系统的多故障诊断。

Study for multiple fault diagnosis technology for spacecrafts based on multi-signal model

The optimal multiple fault diagnosis problem is NP-complete. The multisignal model for complex spacecrafts system such as satellite power was presented which uses diagram to subscribe the causal dependency relationship between components and tests. Consideration for priori probability of the components failure of system and available test set, based on Lagrange relaxation and subgradient optimization, a heuristic algorithm to find the most likely multiple fault set was given. The modeling method and diagnosis algorithm was validated in one test for a certain satellite power system. The results show that the mode is easy to build and the reasoning process is very fast which are suitable to be applied in multiple fault diagnosis for spacecrafts.