The Application of ART2 Two-Factor Optimization Method to Engine Fault Diagnosis

摘要

In order to make ART2 network in engine fault diagnosis of the unknown fault. The vigilance factor and the adjustment factor were proposed to control the re-learning of the known fault. Through judging the credibility of the sample and amending the study mode, adjust the top-down and bottom-up vector. The analysis shows that the two-factor method solved the problem of the lack of training samples and overcomed the disadvantage of the traditional neural network that it was unable to identify the defects. This method can assure the network to continue self-learning and optimization by effectively classifying and identifying the state model of the engine.