摘要
In the fault prediction of mechanical equipments through spectromectric oil analysis for worn off debris, a method for the determination of the limiting value of wear is proposed and discussed. In order to diagnose the impending failure and to predict the fault modes and locate the fault spots, a comprehensive approach is studied and outlined on the basis of methods of discriminative analysis and fuzzy logic. A fault diagnosis expert system OAFDS developed by the authors for the nonitoring of working conditions of the ND5 locomotive diesel engine Nd5 is briefly introduced.
通过光谱油料分析，研究了机械装备磨损金属元素界限值的制定方法．为了识别故障有无、故障部位、故障类型和故障程度，应用数学诊断原理，研究了判别分析法、模糊聚类法和灰色模型预测法．文中以ＮＤＳ柴油机为应用对象，简要介绍了作者开发的ＯＡＦＩＢ故障诊断专家系统．
