Multi-parameter comprehensive diagnosis technology for electrical faults based on artificial intelligence algorithms

This study addresses the poor diagnostic performance of conventional methods in multi-parameter electrical fault diagnosis by introducing a backpropagation neural network (BPNN) model enhanced through genetic algorithms and artificial firefly swarm optimisation. The approach leverages the genetic algorithm's search ability and the firefly algorithm's optimisation capability to improve fault diagnosis accuracy. Experimental results show that the proposed model achieves an average fitness of 95.36%, an average absolute error of 1.837, and an average accuracy of 96.54%, significantly outperforming two commonly used diagnostic models. Performance evaluation based on fitness, loss value, accuracy, and absolute error confirms the model's strong diagnostic capability. The proposed method offers a robust solution for accurate multi-parameter electrical fault diagnosis.