Research and Application of Deep Belief Neural Network for Distribution Transformer Feeding Area Line Loss Reduction

The line loss rate is an important technical and economic indicator for power supply company, which is also an important content to measure company's comprehensive management level and operating efficiency. This paper combines cluster analysis and neural network algorithms to classify distribution transformer feeding area line loss of shanghai power grid. DBN obtains the initial parameters of the network by layer-by-layer unsupervised pre-training, which effectively solves a series of problems caused by the random initialization of traditional neural network parameters. In addition to having a good initial point, DBN can also effectively solve the over-fitting and under-fitting problems that often appear in neural networks through pre-training. This paper verifies the effectiveness and accuracy of DBN model, revealing different loss related electrical characteristics. By learning a large amount of historical data, training the line loss neural network model and obtaining the change rule of distribution transformer feeding area of Shanghai Power Grid, this paper presents a guiding solution for loss reduction measures. For missing operating data input scenarios, this neural network model can also provide loss rate confidence interval.