Dynamic combined prediction method for error state of capacitor voltage transformer based on elastic net regression

The error state of capacitor voltage transformer (CVT) is an important index affecting the fairness of power system transactions, but its long-term stability is poor. In order to grasp the change of CVT error state in time, this paper proposes a dynamic combination forecasting method based on elastic network regression. This method starts with training the historical data of CVT error state by using multiple machine learning models. Then, white noise test are carried out on the trained model to judge the quality of the model. Furthermore, the prediction results of the model that conforms to the white noise test are regarded as input features to construct the elastic network regression model and find the optimal weight coefficient. Finally, the elastic network with the optimal coefficient is utilized for prediction. The experimental results show that each evaluation indexes of the proposed model under single-step prediction is lower than that of each single model, which can effectively improve the prediction accuracy.