Research Of Power Quality Disturbance Detection Based On Improved S Transform

The extensive use of non-linear, impact loads and electronic devices makes the power quality problems become prominent. Power quality issues have gotten more and more attention by power units and customers in recent years. Accurate detection and recognition of power quality is the premise for improving power quality. Power quality disturbances are complex because there are both steady disturbance and transient disturbance. Single time domain or frequency domain analysis is only suitable for few kinds of power quality disturbances, so time-frequency domain analysis becomes an important tool for disturbance analysis. The paper proposes a power quality disturbance detection method based on S transform and power quality disturbance classification method based on improved template similarity.The S transform has been improved by using hyperbolic window instead of Gaussian window in the paper. And modify the hyperbolic window to gain better time and frequency resolution and ability in transient analysis. It combines modified S transform and Dynamic to improve the computing speed and proposes the modified incomplete HS transform method for power quality detection. Simulation results verify that the modified incomplete HS transform is suitable for power quality detection. The features of disturbance extracted from modified incomplete HS transform are more accurate and has lower computation than the traditional S transform. Do modified incomplete HS transform to power quality disturbance, and get the information of prominent frequency points and the module time-frequency matrix. Establish template matrix of the disturbance signal by scale transforming to the corresponding part of the module time-frequency matrix. Then calculate average similarities between template matrix and each standard template. And the type of disturbance can be recognized according to the information of prominent frequency points and similarity. Simulation results show that the proposed classification method has high accuracy rate for both single disturbances and complex disturbances and is robust to noise.