FPGA implementation of edge-side motor fault diagnosis using a Kalman filter-based empirical mode decomposition algorithm

Despite the successful use of deep learning in motor fault diagnosis, its real-time applications have been greatly hindered due to the enormous computational burden and extensive processing time. Addressing this, a Kalman filter-based empirical mode decomposition (KF-EMD) algorithm is proposed, replacing the cubic spline interpolation of traditional EMD with a Kalman filter for real-time FPGA (Field Programmable Gate Array) processing. This algorithm enhances the detection of data anomalies and reduces the computational burden on a lightweight multilayer perceptron (MLP) model, which recognizes features extracted by KF-EMD at fixed intervals. The proposed real-time fault diagnosis method achieved 98.96% accuracy on the Case Western Reserve University (CWRU) dataset and 98.05% on our motor diagnosis experimental platform.