Research On Application Of Rule Reasoning In Fault Diagnosis

The good operation of mechanical equipment is the key to the safety in production and profit for enterprises, which determines an enterprise’s market competition. Fault diagnosis technology, which is important to the normal production of enterprises, ensures that equipments can work safely and efficiently.Fault diagnosis is one technology that is used to monitor the operational status of mechanical equipment by using some detection means to detect whether they are in good order according to data collected and the accumulation of experience. By this way, it can monitor and find the reason for the faults in time, meanwhile predict their trend. The most prominent role of fault diagnosis technology embodies in equipment maintenance system reform, using predictable overhaul instead of the traditional periodic maintenance. Adopting this kind technology, enterprises dispense with some of the excess periodic maintenance operations, reduce corporate maintenance expenses significantly. Predictable overhaul allows enterprises to reduce the maintenance blindness and just target on the fault location for repair, which shortening the repair time and increase the normal equipment operation time. The productivity has been greatly improved and brought additional economic profit. In recent years with a large number of experts and scholars in this field of work, this technology has always been developing and improving. Now, with the research of fault diagnosis expert technology and artificial intelligence neural network, expert system and neural network have been applied to the fault diagnosis, as a result, the efficiency and accuracy of fault diagnosis has been guaranteed.Firstly, this paper studies the rule-based fault diagnosis expert system, and make deep research on the rule matching of inference engine, which is the key part in the rule-based expert system. What’s more it focuses on the analysis and understanding of the Rete algorithm, and puts forward a reasoning machine model. This reasoning machine model includes two parts,one is that is makes relevant improvement from the join operation of the beta network and the relationship between the rules, and the other is that it puts forward a vague-Rete algorithm for the fuzzy rule-based reasoning by adding weight to antecedents. The reasoning machine model is introduced into the fault diagnosis system according to an army equipment fault diagnosis demand, and the basic framework of fault diagnosis system, which is based on improved rete algorithm, is proposed. Finally, the system is designed and implemented.