The invention discloses a method for predicting the concentration of dissolved gas in
transformer oil based on PSO-LSTM. The method achieves the effective evaluation of the operation state of a
transformer through the accurate prediction of the concentration of the dissolved gas in the
transformer oil. The method comprises the steps of firstly, collecting online oil
chromatography sample data of atransformer, determining state characteristic parameters of the data, performing normalization
processing, and dividing a
training set and a
test set; secondly, constructing a long-term and short-
term memory network prediction model, optimizing the long-term and short-
term memory network prediction model through a
particle swarm algorithm to obtain two optimal prediction
model parameters, and reestablishing a long-term and short-
term memory network model according to the obtained optimal prediction
model parameters; and finally, by taking the concentrations of seven characteristic gases dissolved in the oil as inputs and taking the concentration of the gas to be predicted as an output, predicting the concentration of the dissolved gas in the
transformer oil. The method provided by the invention can accurately predict the change of the concentration of the dissolved gas in the
transformer oil, can provide a certain theoretical basis for fault diagnosis and operation
condition evaluation of the transformer, and provides a reference for operation and maintenance personnel to overhaul.