Transient Frequency Response Test and Measurement Error Prediction of DCTV based on AI

Abstract

To solve the problem that the Frequency Response (FR) test scheme and equipment of the traditional DC Voltage transformer (DCTV) used for DC transmission are insufficient, and the Measurement Error (M-E) of the related measurement system is difficult to achieve accurate prediction, a temporary FR test and M-E prediction method of DCTV based on Artificial Intelligence (AI) is proposed. Firstly, the frequency characteristic of DC side voltage of DCTV is analyzed. On this basis, a DCTV transient FR testing method based on transient AC & DC superposition was developed. Then, the method of voltage sudden change and phase correction is used to achieve transient process DCTV response time testing. Finally, the Ant Colony Optimization (ACO) algorithm was improved by combining an adaptive inertia weight improvement strategy, achieving accurate prediction of the M-E of DCTV. The proposed AI based DCTV transient FR testing and M-E prediction method were compared and analyzed with the other three methods through simulation experiments. The results indicate that under the same experimental conditions, the prediction results of the DCTV measurement ratio error and phase angle error of the proposed method are optimal.