Classification Method of Power Quality Disturbance Based on Bidirectional Gating Recurrent Neural Network

Abstract

The invention discloses a method for classifying a single element in a sequence based on a bidirectional gated cyclic neural network. The steps include: 1) manually classifying the collected time series signals or data; 2) converting the input data set and label set into Matrix form; 3) Randomly divide the input data set and the corresponding label set into a training set and a test set, wherein the training set data accounts for 70% of the total samples, and the test set data accounts for 30% of the total samples; 4) Construct a bidirectional gating loop Neural network model; 5) training the constructed bidirectional gated cyclic neural network model; 6) overfitting judgment; 7) using the trained bidirectional gated cyclic neural network model to classify the single element of the sequence, using the Argmax function The final judgment result is obtained from the output layer, and the correct classification of a single element in the sequence is obtained. With the method of the invention, the recognition accuracy rate of the sequence data is over 99%.

Description

technical field [0001] The invention belongs to the technical field of signal control, and relates to a single element classification method in a sequence based on a bidirectional gating cycle neural network. Background technique [0002] Classifying single elements in the sequence of waveforms or data with time-series characteristics is widely used in practical engineering, such as monitoring the operating status of the power grid, monitoring the operating parameters of equipment, identifying the type of power quality disturbance, and identifying vibration signals , the recognition of ECG signals, the recognition of audio signals, the attribute judgment of seismic wave types, etc. [0003] The current solution to the classification of time series signals is usually to determine the category corresponding to the data after monitoring the data for a period of time. It is difficult to classify the single element information in time series signals or data, so the real-time poo...

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to provide a single element classification method in a sequence based on a bidirectional gated cyclic neural network, which solves the problem in the prior art that cannot efficiently and accurately classify and identify a single element in sequence data

Images