Research on Commutation Failure Fault Diagnosis in HVDC Transmission Systems Based on WGAN and DT-SVM

Commutation failure is a typical fault in high-voltage direct current (HVDC) transmission systems. It can lead to drastic changes in a series of electrical parameters, such as the voltage on the inverter side and the direct current, significantly impacting the safety and stability of the receiving AC system. Therefore, accurate and timely diagnosis of commutation failure is crucial for implementing subsequent control measures to prevent maloperation of protection devices. To tackle the difficulty in accurately diagnosing commutation failure faults under the influence of multiple factors, a fault diagnosis method based on a combination of Wasserstein Generative Adversarial Network (WGAN) and Decision Tree Support Vector Machine (DT-SVM) is proposed. Firstly, the WGAN is employed to augment the sample data of commutation failure faults. Then, a DT-SVM-based fault diagnosis model is constructed. Finally, the augmented sample data is analyzed for fault diagnosis. Experimental results demonstrate that the proposed method can accurately identify the types of commutation failure faults in HVDC transmission systems.