Analysis of class group distinguishing based conceptual models for multiple fault diagnosis

It is common that multiple fault exists in actual engineering and complex systems. Due to parameters in multiple faults tightly coupled, relationship between the fault mode and known mono-fault features is non-linear. Thus, it is hard to see how distinguish in mapping set for "fault to symptom". In this case, there is no guarantee that traditional diagnosis methods for mono-fault meet the demands. With the requirement, an analysis of the traits of multiple faults is made. A summarization is given to class group distinguishing (CGD) based methods that applied in fault diagnosis. Major defects in the methods that applied in multiple fault diagnosis are analyzed. On that basis, fault modes and symptoms are taken as key points. Conceptual models for multiple fault diagnosis based on CGD are gradually explored. By the models, actual faults can be mapped to one or more known mono-faults via distinguishing analysis, and therefore multiple faults can be diagnosed. There are 4 kinds of flow chart and construction for the models are established. Each of these models presents advantages and disadvantages are separately presented at the end of the chapter.