Fault diagnosis method based on improved PSO-FCM-immune algorithm

Because of the deficiency that non-consistence of data amount with one fault type of the historical fault data of hoisting machine can cause the diagnose accuracy to be unstable, this paper proposes a fault diagnosing method of hoisting machine based on particle group clustering immune algorithm. This method firstly makes use of an improved FCM clustering algorithm from PSO to classify historical fault data of hoisting machine, and then enriches every type of fault data with particle group and immune algorithms to generate different fault diagnosing machines of different fault types which can be used to do fault diagnosis. Finally, this paper does the experiments to demonstrate that the improved algorithm overcomes the defect of the former algorithm, improves the accuracy of fault diagnosis, and gets better results of fault diagnosis. Copyright © 2013 Binary Information Press.