Research On Interence Algorithm Of Expert System Of Fault Diagnosis For Large-sized Diaphragm Pump

This is a practical subject co-developed by Northeastern University Equipment Diagnosis Center and Anshan Iron and Steel Group Corporation to meet the need of monitoring and testing the high-power reciprocating diaphragmatic pump.The reciprocating diaphragm pump is one of the most important machines in the field of industry manufacturing, its state of running is directly related to the efficiency of producing, thus monitoring its running state is very important, in order to make trend prediction of the Machine and maintaining in time and avoiding maintaining too often or not enough. With the development of computer, it is a new way of diagnosis that theory and method of Artificial Intelligence is applied to fault diagnosis and intelligent diagnosis technology is developed. It is a main direction that intelligent diagnosis technology has been applied extensively.Moreover, as the important portion of the diaphragm pump, the hydraulic power system and valve's fault is very hackneyed, and its fault can cause big expenses to producing. Therefore, the research work of this article is about the fault of the valve and hydraulic power system. It also gives out the diagnosis result. The main research work and its production are as follows:(1) Based on inference to relative document, allow for character of knowledge of field, object-oriented Paradigm is adopted and a knowledge base and its knowledge base management system are built.(2) According to character of knowledge of field, uncertainties inference is adopted in expert system. Moreover, based on inference to relative document, fault tree of the hydraulic power system of diaphragm pump is built and parameters of uncertainties are calculated, using the transferable belief model.(3) Five kinds of training function are discussed. Allow for portability and feasibility of expert system. Interface technology is adopted according to different training function. According to character of knowledge of field, combining with fuzzy mathematics, method of fuzz judgment is given to output of ANN.(4) Allow for character of knowledge of field of diagram pump, flow of work of inference engine is given. Moreover, integral construction is given and interface technology is discussed.(5) Combining with knowledge of field of valves of fault, example of diagnosis of valves is given.