抄録
Condenser fault diagnosis model is improved in the paper by using the method of fuzzy math and the technology of ANN. Typically fault-sign sample of condenser is redefined in the paper. The change of condenser parameters is not always caused by fault. It may be caused by the variety of unit load and environmental factors. So it is necessary to eliminate objective factors in fault diagnosis. The paper reestablished membership functions of running parameters that are changed according to unit load and environmental factors. New training arithmetic is used to shorten the time of diagnosing condenser fault. The structure of condenser fault diagnosis system is provided. The result, which is got by using the system, is consistent with the one that is deduced by using operating data and experiences.
