FD-ELM Diagnosis Method for Electric Actuator of Six-Degree-of-Freedom Motion Platform

The existing fault diagnosis methods based on the neural network have some problems, such as inaccurate diagnosis positioning and insufficient manual feature extraction in the variable working conditions of the electric actuator of the six-degree-of-freedom motion platform. To solve these problems, this paper proposed a diagnosis method of the electric actuator based on the fixed dictionary-extreme learning machine (FD-ELM):1. The variational mode decomposition (VMD) decomposed the sensor signals.2.The singular value decomposition (SVD) algorithm extracted the features of the modes decomposed by VMD to weaken the influence of noise.3.FD-ELM adaptive to variable speed is used to learn the extracted features, and the learned FD-ELM algorithm model completes the fault identification of the electric actuator. Experimental results show that the proposed method can effectively identify the faults of the electric actuator under varying working conditions. In addition, the results of the proposed method are better than the existing neural network fault diagnosis methods in the comparison experiment based on the vibration data set of the real six-degree-of-freedom motion platform of the full-function train driving simulator.