Loss prediction model of distribution network based on quantitative data calculation

In order to improve the accuracy of distribution network loss prediction model, this paper proposes a loss prediction model of distribution network based on quantitative data calculation. Firstly, the load imbalance coefficient, which represents the degree of system load imbalance, is defined and combined with the theoretical calculation of network loss to simulate the generation of system network loss numerical data; Then, the correlation between load level, imbalance level and total network loss is qualitatively analyzed, and a deep confidence network (DBN) is constructed to learn the sample data and obtain the implicit representation of the correlation between characteristic quantities. The simulation results show that DBN can effectively describe the complex mapping relationship between load imbalance coefficient, load apparent power and total network loss, and can quickly and accurately predict the network loss of distribution network.