Method of identifying and classifying transient electric energy quality recording data

Abstract

The invention discloses a method of identifying and classifying transient electric energy quality recording data. The method includes the following steps: obtaining electric energy quality recording data, and screening training data and test data among the data; converting the data into a module time frequency matrix, and extracting a feature vector of the module time frequency matrix; according to the feature vector, establishing a classifier of the electric energy quality recording data on the basis of a BP neural network; establishing a training sample and a test sample; inputting the training sample into the classifier, optimizing the BP neural network of the classifier through a PSO algorithm, and obtaining an optimized classifier of the electric energy quality recording data on the basis of a PSO-BP neural network; inputting a test feature vector into the optimized classifier, and receiving test classification output from the optimized classifier; determining whether the test classification and expected test classification are consistent; and if the test classification and the expected test classification are consistent, outputting the test classification. The method has the advantages of being high in identification efficiency, being high in identification accuracy, and exhibiting the high anti-interference capability.

Description

technical field [0001] The invention relates to the technical field of power quality analysis, in particular to a method for identifying and classifying transient power quality wave recording data. Background technique [0002] With the rapid development of science and technology and the national economy, new energy sources such as photovoltaics and wind power in the power system are connected to the grid on a large scale. In addition, more and more large-capacity nonlinear loads are widely used in the power system, such as electrified railways, metallurgical etc., making the power quality problem of the power system more and more serious. In order to grasp the impact of power quality problems on production activities, power quality monitoring points are set up at many grid-connected points of nonlinear loads to realize continuous monitoring of power quality at grid-connected points. The power quality monitoring device can not only obtain steady-state power quality data und...

AI Technical Summary

Problems solved by technology

Due to the problems of spectrum leakage, susceptibility to noise, and lack of intuition of the transformation results in wavelet transform, and the input data is complex and huge, the training and recognition of neural network is difficult, the training speed is slow, and the training accuracy is low, which ultimately affects the whole method. Recognition and Classification Effects

Images