A composite power quality disturbance identification method and method based on energy maximization and a kernel SVM

Abstract

The invention discloses a composite power quality disturbance identification method based on energy maximization and a nuclear SVM. The method gets rid of the complexity of manual design characteristics, simplifies electric energy quality classification into two steps, and comprises the following specific steps: 1) performing time-frequency analysis on composite electric energy quality disturbanceby adopting an S transformation method based on energy maximization, automatically adjusting window parameters to realize optimal time-frequency resolution; and 2) establishing a weight linear combination kernel function in combination with the extracted time-frequency characteristics to form a kernel SVM algorithm. The kernel SVM can fuse various features, and each feature corresponds to the optimal kernel function parameter, so that the kernel SVM is more adaptive to the disturbance of the composite power quality. The defects of time consumption and information loss of manual features are overcome, and the accuracy of power grid signal recognition under complex working conditions can be further improved.

Description

technical field [0001] This patent belongs to the field of power quality detection and classification, and in particular relates to a composite power quality disturbance identification method and method based on energy maximization and kernel SVM. Background technique [0002] With the increasing use of nonlinear disturbance loads in power systems, the types of power quality signals are becoming increasingly complex and diverse. Common non-linear loads include charging piles, power electronic switches, etc. In addition, the grid connection of some clean energy sources, such as wind energy, solar energy, and geothermal energy, also increases the need for power system signal governance. In order to provide users and enterprises with cleaner energy, the country spends a lot of money and manpower on the regulation and governance of power quality signals every year. Based on this, in order to ensure more accurate management and detection of power system signals, the first prereq...

AI Technical Summary

Problems solved by technology

The artificial neural network has strong fitting ability, but the parameters have a great influence on the classification

Among these classification methods, SVM has a complete and rigorous mathematical reasoning and proof process, and its classification effect is interpretable, but the single kernel function in ordinary SVM is difficult to meet the multi-dimensional feature classification requirements.

Images