Real-time FPGA classification of power quality disturbances using modified S-transform and RBF neural networks

— Power quality (PQ) disturbances are increasingly affecting modern power systems due to the widespread use of nonlinear loads, renewable energy sources, and power electronic converters. Reliable and real-time classification of PQ disturbances is essential to ensure power system stability and equipment protection. This paper presents an FPGA-based real-time PQ disturbance classification method using a Modified S-Transform (MST) for feature extraction and a Radial Basis Function Neural Network (RBFNN) for classification. The proposed MST improves time–frequency resolution while reducing computational complexity, making it suitable for hardware implementation. Discriminative features extracted from the MST are classified using an RBFNN due to its fast convergence and high classification accuracy. The complete system is implemented on an FPGA platform to achieve real-time performance. Simulation and hardware results demonstrate high classification accuracy for various PQ disturbances, including voltage sag, swell, interruption, harmonics, flicker, and transients, with low latency and efficient hardware utilization.