Study On Fault Diagnosis Method Of Turbine Vibration Based On Pattern Recognition

Combining the negative selection principle of immune algorithm which was introduced in fault diagnose of turbine rotator vibration and the rapidness, accuracy belongs to immune algorithm application in pattern recognition, we developed an improved approach on antibody caused by vaccine. The method-featured by inspection parallel with self-learning and successively dynamic regulation, was implemented by reasonable antibody set attributing to regular working sample and a few of faults sample, which result in rapid and precise recognition. Consequently, series of vibration data was thus been obtained, furthermore, three categories of characteristic parameters was adopted and base thereon the algorithm's validity was accordingly verified.