Fault Diagnosis of Subway Plug Door Based on Isomap and GWO-SVM

In order to solve the problem of high false alarm and missing alarm rate in health fault diagnosis of subway plug door, this paper proposes a health fault diagnosis model of subway plug door based on Isomap and GWO (grey Wolf swarm optimization support vector machine (SVM). According to the nonlinear and non-stationary characteristics of the current curve of the subway plug door motor, the model constructs multi-feature parameters, Uses Isomap dimension reduction optimization, and inputs the feature set after dimension reduction into the support vector machine based on gray Wolf swarm optimization for training. Finally used in subway plug door experiments have been carried out to verify the measured data, the results show that support vector machine (SVM) based on grey Wolf group algorithm optimization diagnosis model than directly using the BP neural network or support vector machine (SVM) diagnosis model has higher precision of fault diagnosis, relative experimental error smaller and has certain actual application value.