Fault Detection And Operation Optimization Of Smart Microgrid Based On Deep Learning

With the rapid development of new energy power generation technology and communication technology,the market demand and input-output ratio of smart microgrids are getting higher and higher.Smart microgrid integrates the advantages of smart grid and microgrid.It has a high degree of automation,flexible and reliable power supply,and is a good carrier for studying smart grids and their operation control methods.Compared with the traditional power grid,the smart microgrid has a complex structure,and there are many potential faults that are difficult to detect by traditional methods,resulting in changes in the configuration structure,which will have a more obvious impact on the safe and stable operation of the smart microgrid,and the efficiency of the system will decrease or even be systematic Instability.This paper combines smart microgrid AI fault detection with network optimization and reconstruction technology and hierarchical control strategy to improve the operating efficiency and stability of the smart microgrid under fault conditions.First,a smart microgrid model containing three energy forms of cold,heat and electricity and smart regulating equipment is constructed,and a micro-source simulation model is established.Next,a smart microgrid hierarchical control architecture based on fault identification and network reconfiguration is designed.The principle of the droop control strategy of the bottom-level grid-connected inverter is analyzed,and the bottom-level droop control simulation module of each micro-source is constructed.The automatic power distribution function of droop control has been simulated and verified.Secondly,in view of the volatility of distributed power sources and loads,aiming at operating costs,the particle swarm algorithm is used to optimize the power setting of the micro-source grid-connected inverter,reducing the operating cost of the micro-source of the smart micro-grid and increasing energy utilization rate.Aiming at the problem of smart microgrid fault detection,deep learning algorithms are used to build a smart microgrid AI fault recognition and location system,and a convolutional neural network structure is used to extract simulation data under different states of the smart microgrid,and a training set is constructed.Through the inspection and network optimization,fault identification classification and location have been completed.Finally,for the problem of intelligent microgrid re-optimization in the fault state,the network reconstruction technology is combined with the power optimization algorithm,and the power flow calculations are carried out one by one for various possible network structures in the fault state,and the fast traversal algorithm is used for different operating conditions.The smart micro-grid of the United States carries out network optimization and reconstruction,comprehensively optimizes the operation of the smart micro-grid,and improves the overall economy and energy efficiency of the system.A simulation platform for smart microgrid,fault identification network and hierarchical control system under the Matlab platform was constructed.For the AI fault detection and location system based on deep learning algorithms,30% of the training data set was selected and various fault identification verifications were performed.The simulation results show that the recognition accuracy of explicit faults can reach 99.9%;a multi-condition operation optimization model based on the upper-level power optimization control strategy and the network optimization reconfiguration algorithm is constructed,and the simulation verification is carried out in combination with the lower-level droop control.The simulation results show :The optimization strategy reduces the line loss cost and the micro-source operating cost,so that the overall operating economy of the smart micro-grid is optimal.