Multi-Agent Fault Diagnosis in Manufacturing Systems Using Soft Computing

The expeditious and accurate diagnosis of faults in manufacturing systems is essential in order to avoid expensive downtime. Many artificial intelligence approaches to automated fault diagnosis use techniques that are too computationally complex to achieve a diagnosis in real-time or are too inflexible for dynamic systems. Other approaches use either structural or symptom-based reasoning. Functional approaches are unable to provide real-time response due to their computational complexity, whereas, symptom-based approaches are only able to handle situations specifically coded in rules. Current hybrid approaches that combine the two methods are too structured in their approach to switching between the reasoning methods and, therefore fail to provide the flexible, rapid response of humans experts. This paper presents a robust, extensible approach to fault diagnosis that allows unstructured switching between reasoning methods using multiple fuzzy intelligent agents that examine the problem domain from a variety of perspectives.