Abstract:

Since the supercritical generating system has complex structure and high requirements on safety performance, fault diagnosis is very difficult. Combining fuzzy theory and DS evidential theory, a new fault diagnosis method is proposed, whose process includes membership function determination, fuzzy processing, neural network diagnosis, data fusion with DS evidence theory and obtaining the final result. Taking the fan of a supercritical generating system as an example, this method can be used for fault diagnosis, finding the equipment with higher fault probability quickly and accurately and providing countermeasure according to the fault in time.

Key words:

supercritical unit, fault diagnosis, fuzzy algorithm, DS evidential theory, data fusion