Power Quality Disturbance Recognition Based on S-Transform and SOM Neural Network

This paper presents a new approach for recognition of nonstationary signal in power quality (PQ) disturbance using S-transform and self-organizing mapping (SOM) neural networks. The most common types of the PQ disturbance, such as voltage sags, swells, interruptions, transients and harmonics, are studied. We utilize S-transform to process power quality disturbance signal. In this step, the obtained features can overcome the noise of the original nonstationary signals. Meanwhile self-organizing mapping (SOM) neural networks are utilized to solve PQ disturbances classification from the obtained features. The simulation result shows the validity and feasibility of the proposed model.