基于语义推理和数据融合的IETM综合诊断系统

IETM integrated diagnosis system based on semantic inference and data fusion

为满足复杂装备对故障诊断的需求,面向智能交互式电子技术手册(interactive electronic technical manual,IETM),提出了结合模糊语义推理和贝叶斯信念融合方法的综合诊断系统。对清晰的故障征兆,首先通过基于经验的人工判读进行直接诊断;同时利用IETM的模糊语义推理结合故障征兆开展快速诊断。对于核心在线监测部件或子系统,采用基于状态的智能维护,利用贝叶斯算法融合多分类器决策进行增强诊断。以动车组空气压缩机和牵引电机为例,实验验证了该综合诊断系统能够应对不同复杂程度的诊断需求,诊断准确且快速有效。

In order to satisfy the demand of complex equipment for fault diagnosis, facing intelligent interactive electronic technical manual （IETM） , this paper presents a novel integrated diagnosis system, which integrates experience-based manual interpretation, fuzzy semantic inference and Bayesian belief fusion methods. For the clear fault symptoms, the manual interpretation based on the experience of onsite maintainer is adopted to conduct direct diagnosis ; meanwhile, the fuzzy semantic inference exerted on IETM is combined with the fault symptoms and used to conduct rapid diagnosis. In terms of core on-line monitored components or subsystems, the condition- based intelligent maintenance is adopted, the Bayesian belief algorithm is fused with multiple classifier decision and used to conduct enhanced diagnosis. Taking the air compressor and pulling motor of electric multiple units （EMU） as examples, fault diagnosis experiments were conducted to evaluate the developed system. Experiment results verify that the developed integrated diagnosis system can deal with the diagnosis requirements with different fault complexity, and achieve fast, effective and accurate fault diagnosis performance.