Prediction method for dissolved gas concentration in transformer oil based on long and short-term memory network

Abstract

The invention discloses a prediction method for dissolved gas concentration in transformer oil based on a long and short-term memory network, which realizes assessment of a transformer running state through accurately predicting concentration dissolved gas in transformer oil. The method comprises the following steps: firstly, collecting transformer online oil chromatographic sample data, determining characteristic parameters of the data and performing normalization; then, training the long and short-term memory network with sequence data of the dissolved gas in the transformer oil, and obtaining an optimal prediction model parameter; finally, taking seven types of characteristic gas concentration dissolved in the oil as input, taking concentration of the to-be-predicted gas as output, andrealizing prediction of the dissolved gas concentration in the transformer oil. The method provided by the invention can accurately predict change of the dissolved gas concentration in the oil, provides basis for judgement of running conditions of a power transformer, and provides reference for operation and maintenance staffs to perform overhaul.

Description

technical field [0001] The invention relates to the technical field of power equipment monitoring, in particular to a method for predicting the concentration of dissolved gas in transformer oil. Background technique [0002] As the most core equipment in the power system, the power transformer plays an important role in the distribution and transmission of electric energy. It is an important asset of the power grid company. Its safe and stable operation is the prerequisite for ensuring the reliable power supply of the power grid. Under normal operation of the transformer, due to the aging of the internal insulating solid, a small amount of gas will be dissolved in the insulating oil, mainly hydrogen (H 2 ), methane (CH 4 ), ethane (C 2 h 6 ), ethylene (C 2 h 4 ), acetylene (C 2 h 2 ), carbon monoxide (CO), carbon dioxide (CO 2 ) and other gases. Different operating conditions of the transformer can be judged according to the concentration ratio of the dissolved gas ...

AI Technical Summary

Problems solved by technology

Due to the disadvantages of the traditional back propagation neural network model (Back Propagation Neural Network, BPNN) and support vector machine (support vector machine, SVM) prediction methods have the disadvantages of long period, large error, slow convergence speed, and different network structure selection, etc., cannot Predict and analyze the concentration of dissolved gas in time, so it will delay the judgment of the operation status of the transformer, resulting in economic losses

Images