Under the guidance of the concept of green development,new clean energy is the main direction of power generation in the future.However,after years of construction,China’s thermal power generation and supporting power grid system has been relatively mature,the existing distribution network can not meet the access of a large number of distributed generation,which limits the development of new clean energy.In this context,the concept of active distribution network is proposed,but in order to achieve the optimal control effect,the control method with good performance is the core of active distribution network.Therefore,this thesis studies the control method and application of active distribution network based on neural network.In order to fully understand the active distribution network,the structure and characteristics of active distribution network are analyzed in detail.On this basis,the specific problems faced by the active distribution network are analyzed,and it is considered that the active distribution network needs to have a good performance control method in order to maximize the effectiveness.In order to realize the optimal control,this thesis analyzes the centralized,hybrid and hybrid hierarchical control structures,and finally selects the hybrid hierarchical control structure.This thesis analyzes the feasibility of neural network intelligent algorithm as the core control method of active distribution network,and finally selects multilayer convolution neural network as the core algorithm.In this thesis,the multi-layer convolution neural network algorithm is analyzed,and it is considered that the accuracy and loss value of the algorithm in the training process do not meet the expected requirements,so the determined method gradient algorithm is used to improve it.Through simulation analysis,the accuracy and loss value performance of the improved algorithm are greatly improved.On the basis of theoretical research,an active distribution network control model with node voltage and power parameters as optimal control objectives is established.In order to verify the performance of the control model,the IEEE 18 bus distribution network simulation model is established.Through the simulation and analysis of the voltage deviation of the distribution network nodes,the proposed active distribution network control model can meet the requirements.In the end,the control method of active distribution network is applied based on the actual project.The stability of distribution network and the rate of distributed generation are analyzed.The results show that the control method proposed in this thesis can meet the needs of practical application. |