Early Intelligent Fault Diagnosis of Rotating Machinery Based on IWOA-VMD and DMKELM

Abstract

Abstract The effect of early fault vibration signals from rotating machinery is weak and easily drowned out by intensive noise. Thus, it is still a great challenge to perform early fault diagnosis. An intelligent early fault diagnosis method for rotating machines is proposed based on Variational Mode Decomposition (VMD) parameter optimization and Deep Multi-Kernel Extreme Learning Machine (DMKELM). First, the iterative chaotic mapping, nonlinear convergence factor and inertia weight are introduced into the whale optimization algorithm (WOA). Then, an improved WOA (IWOA) is presented and utilized to optimize VMD parameters. The optimized VMD (OVMD) combined with the sampling entropy is used to denoise and reconstruct the signals. Finally, the mish function is served as the activation function. Meanwhile, the radial basis kernel function (RBF) and polynomial kernel (PK) are introduced to construct the mixed kernel function. The mixed kernel assignment is applied to replace the random assignment, and the DMKELM is presented to enhance the classification performance and generalization capability of the model, and DMKELM is used for early intelligent fault diagnosis. Two experiments on bearings and gears demonstrated that the fault diagnosis accuracy is 99% and 98.5%, respectively, and the fault diagnosis accuracy is high. It shows that this method has great superiority in early fault diagnosis of rotating machinery.