Research on Prediction Model of Gas Concentration in Power Transformer Oil

The research of gas concentration prediction model in power transformer oil is an important means to diagnose the fault type of power transformer. According to the gas concentration generated in the daily work of the power transformer, the gas concentration generated in the later power transformer oil can be predicted. According to the predicted gas concentration, the fault type of the power transformer can be better diagnosed in advance. Therefore, the prediction of gas concentration in power transformer oil is discussed and studied. For the gas prediction of power transformer oil, long short-term memory network prediction model, bidirectional long short-term memory network prediction model and gated recurrent unit prediction model are selected. Taking the hydrogen concentration generated by a power transformer as an example, experiments are carried out, and R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> and RMSE of the three models are compared and analyzed to evaluate which prediction model has better prediction effect.