基于状态分类评价的发动机故障诊断专家系统

Expert System for Fault Diagnosis of Engines Using State Classification Appraise

传统的专家系统,完全在假设的情况下诊断,诊断效率不高,对监控数据很难量化,不易提供线上诊断。为了解决这个问题,提出基于状态分类评价的专家系统,依据发动机构造的特点,将技术状态分成2类:物理类和逻辑类。物理状态通过基本功能参数与零件几何特性关系建立动态故障模型,逻辑状态通过推理对专家知识采用D-S算法和预测故障程度相结合的方法。从而达到对发动机故障进行整体诊断。

The efficiency isn't high in the process of the traditional expert diagnose system,for it was done completely on an assumption and beyond possibility to monitor what it should more exact referring to a quantun,and it is harder to be diagnosed through a network.To solve this problem, the researcher puts forward an expert diagnose system based on a classified state appraisal. According to the specialty of structure of engines,technical conditions are classified into two categories:physics type and logic type. In connection with the fundamental,functional parameters and the geometry relationship of mechanical parts,a dynamic fault model is established,while logic type is set up by adoption of experts' rational knowledge and of using D-S calculation method. In this way, engine faultsdiagnoses could be achieved.