Fault Diagnose of Rotating System Based on ICA with Reference and RBF Networks

This paper presents the technique of fault diagnosis using independent component analysis (ICA) and demonstrates applications of ICA-based RBF networking in the diagnostic system. The ICA with reference is proposed to incorporate additional requirements and prior information as constraints into the ICA constraints into the ICA contrast function. The adaptive solutions using the RBF network learning are proposed to solve the constrained optimization problem. A radial-basis-function (RBF) neural network based fault detection method is developed. The application illustrate the versatility of the method of the paper by separating the subspace of independent components according to density types and extracting a set of desired sources when rough templates are available. The experiments using an unbalance rotor of rotating systems demonstrate the efficacy of the method