The Remaining Life Prediction of the Fan Bearing Based on Genetic Algorithm and Multi-Parameter Support Vector Machine

Abstract

In order to meet the needs of the fault prediction and health management of the cooling fan in the industrial field, a new method based on genetic algorithm and multi-parameter support vector machine (GSVM) is proposed to predict the remaining life of the fan bearing. In this method, the fan system is used as the research object to predict the bearing life, which is closer to the actual working condition of the bearing compared with the research on the bearing. The multi-parameter is used as the input value of the support vector machine (SVM), which is more accurate than the single parameter. Genetic algorithm is used to optimize the parameters of SVM model, which improves the accuracy and certainty of parameter selection. The correlation analysis is used to choose the characteristics that are closely related to the remaining life. The unbalanced load is put on the fan blades to accelerate bearing life prediction process. The result shows that the model established by the parameters of the fan current, voltage and speed can accurately predict the residual life of rolling bearing, which not only provides theoretical guidance for the maintenance and replacement of the fan bearing, but also improves the fan predictive maintenance level.