Power line impulse noise modeling method based on generative adversarial neural network

The invention relates to a power line impulse noise modeling method based on a generative adversarial neural network, and the method comprises the steps: firstly, employing an FPGA to obtain actually-measured power line impulse noise data, carrying out the processing through MATLAB, and obtaining an actually-measured waveform, the noise amplitude of which changes with the number of sample points, and secondly, carrying out the modeling of the impulse noise of a power line. A generative adversarial neural network is adopted to replace a traditional neural network to model power line impulse noise, the network training process is optimized, the training method is more concise, uncertain factors in the training process are reduced, and the convergence speed of the neural network is increased. Besides, a generator and a discriminator in the generative adversarial neural network are continuously trained in high-proportion power electronic equipment according to an analysis result, so that the method can better ada