Study on Power Transformer Fault Diagnosis Based on Niche Genetic Algorithm

Power transformer fault diagnosis is the key technology of electric power system. Niche genetic algorithm (NGA) was introduced to optimize adaptive-learning-rate-momentum back propagation (BP) network. NGA-BP model was established and applied to power transformer fault diagnosis. Compared with BP network, NGA-BP network has a better performance on convergent speed and stability. The experimental results demonstrate that fault diagnosis precision of NGA-BP network is 90%, which is higher than BP network.