A novel Fault Diagnosis Method for Power Transmission Lines Based on Improved Deep Belief Network

With the rapid development of artificial intelligence technology, the application of deep neural networks to power transmission line fault classification is gaining more and more attention from researchers. In this paper, a novel power transmission line fault classification method based on an improved Deep Belief Network is proposed. The Adam algorithm is used to adjust the learning rate. The convergence rate can be faster and the probability of falling into a local extremum is reduced. This paper uses IEEE 30-bus system as an example for simulation experiment. The electrical parameters and their zero components at the key buses of the power system are selected as eigenvalues. Experimental results show that, contrary to conventional Deep Belief Network method, the improved method has higher classification accuracy and faster convergence. When we increase the noise level of the dataset, the improved Deep Belief Network still has better classification effect.