Real-Time Monitoring Method for Thyristor Losses in Ultra High Voltage Converter Station Based on Wavelet Optimized GA-BP Neural Network

As the core equipment for AC/DC conversion in ultra-high voltage direct current (UHVDC) transmission systems, thyristor converter valves are the main source of losses in converter stations. However, it is difficult to directly measure the actual thyristor losses in UHVDC converter stations, and the existing loss calculation methods have many shortcomings, lacking accuracy and real-time performance. In this paper, a real-time monitoring method for thyristor losses in UHVDC stations based on wavelet optimized genetic algorithm-backpropagation (GA-BP) neural network is proposed. Firstly, wavelet transform is used to remove high-frequency noise from thyristor test data and extract features from the original signal. Then, genetic algorithm is used to optimize the initial weights and biases of the BP neural network, and a loss calculation model is constructed through dataset training. Finally, combined with the electromagnetic transient operating point, real-time monitoring of thyristor losses is achieved. Through PSCAD-MATLAB interactive interface simulation verification, this method can obtain real-time power consumption curves of thyristors based on changes in operating conditions. Moreover, compared to traditional fitting algorithms and standard neural networks, the wavelet optimized GA-BP neural network has the advantages of fewer iterations and higher fitting accuracy.