基于混合模型的多重故障诊断

Multiple Failure Diagnosis Based on Mixed Models

介绍了基于模型的多重故障诊断，它可通过ＡＴＭＳ来实现，能克服基于规则的诊断系统的一些局限性，但当运行模型和故障模型同时被运用或者等级模型被运用时，普通ＡＴＭＳ不能处理这种情况．本文应用分层ＡＴＭＳ来实现基于混合模型的故障诊断方法，同时给出广义最小诊断的定义，从而克服关于最小诊断定义中的不足．由于分层ＡＴＭＳ能表示和处理论据之间的优先级，因而能运用混合模型，能结合基于模型和基于规则的诊断，及提高诊断效率．

This paper presents a model based multiple failure diagnosis which can overcome some of the limitations of rule based diagnostic systems. ATMS can be used to realize model based multiple failure diagnosis. But when both correct models and fault models are used or hierarchical models are used, the basic ATMS can't deal with them. This paper uses stratified ATMS to realize multiple failure diagnosis on the basis of mixed models, and gives the definition of general minimal diagnosis which can overcome the inadequacy in the definition of minimal diagnosis. As the stratified ATMS can represent and deal with priorities among justification, it is able to use mixed models, to combine model based and rule based diagnosis, and to improve the efficiency of diagnosis.