Study of fault diagnosis method for three-phase high power factor rectifier based on PSO-LSSVM algorithm

The least squares support vector machine (LSSVM) use quadratic loss function to replace the non-sensitive loss function and equality constraints to replace inequality constraints. LSSVM is widely used in pattern recognition and function regression, but its performance mainly depends on the parameters selection of it. Kernel parameter selection is very important, and which decide the fault diagnosis precision. In order to enhance fault diagnosis precision for electric equipment, the LSSVM algorithm based on PSO is proposed. The algorithm, which can complete automatic parameter selection, is used to choose sigma parameter of kernel function. The experiments show that the PSO-LSSVM algorithm has better fault diagnosis ability than LSSVM.