Prediction of dissolved gases in power transformer oil based on RBF-LSSVM regression and imperialist competition algorithm

In order to accurately predict the dissolved gas content for the transformer oil in a period of time in the future, imperialist competitive algorithm (ICA) and the least squares support vector machine regression (LSSVR) were introduced to the prediction model. ICA is employed to optimize the hyper-parameters of constructed SVM regression, and the parameters of the radial basis function (RBF) kernel function are optimized and the prediction model is established. The results show that the proposed model has obvious advantages and the prediction effect is generally higher than BPNN, which verifies the correctness of the proposed method and the feasibility of the scheme. The MAPE in training of RBF — LSSVR less than 1.3% while BPNN's higher than 8.8%, at the same time, the MAPE in testing of RBF-LSSVR less than 1.8% while BPNN's higher than 9.0%.