A Method for Predicting Power Loss of HVDC Converters Based on Support Vector Regression

Abstract

For a HVDC converter station, the commonly used power loss determination methods are difficult to accurately reflect the changes of power loss of the converter in real time, given that the operating parameters of the converter station are dynamically changing when the converter station is running normally. Therefore, this paper proposes a method for predicting power loss of HVDC converters based on support vector regression. According to this method, firstly, the power loss data of a converter is analyzed. Then the appropriate feature in the power loss data is selected and thus a dataset of power loss samples can be obtained for further work. By applying the support vector regression algorithm to the dataset collected before, it is possible to predict the power loss of a converter for various operating parameters of the HVDC converter station. Finally, the cross-validation method was used to validate the stability of the prediction method. The result of the validation shows that the proposed method is able to accurately and stably predict the power loss of a converter of the HVDC converter station in real time.