Decision And Recognition Of The Power Quality Disturbance Quantity

Power electronic devices, widely used in power systems, lead to a great number of powerquality problems. Those affect the safety and stability of power grid operation and bring aboutserious economic losses. It is particularly important to improve and enhance the power quality.Classification and recognition of power quality disturbance signal is an important part of thepower quality disturbance monitoring automatically, is also a prerequisite of determining thereasons and improving power quality disturbances. The power parameters of the noisedisturbance signal are analysised and the disturbance quantities of influencing power quality arerecognized and classified, which provide a scientific basis for improving power quality.Firstly, based on the principle of mathematical morphology, a multi-structure and adaptiveweights morphological filter with multiple structuring elements is designed to denoise.Secondly, S-transform is applied to perform time-frequency analysis on the denoised signalsand achieve the feature curves in order to analyze the power quality disturbances. Main jobsinclude extracting some characteristic curve such as11times the fundamental frequency curve,time-amplitude envelope curve and frequency-amplitude envelope curve. The characteristicquantity of output curves is extracted and inputted to the probabilistic neural network to classifyand identify disturbance signals. Finally, the two methods are used to identify interferencesources of voltage sags and simulations are focused on voltage sags.Simulation results show that the characteristics of the disturbance signal are retained welldue to the morphological filter; the S-transform has favorable time-frequency analysiscapability and accomplish more accurate detection for disturbance signals; The way ofcombining the S transform and the probabilistic neural network has the speed of training, thegood effect of classification and recognition.