Research On Capacity Prediction Model And Optimal Control Of Energy Storage Device Based On Deep Learning

Deep learning is a research hotspot in the field of artificial intelligence.It has developed rapidly in recent years,and its presence can be seen in many fields.With its powerful automatic feature extraction capabilities,deep learning has powerful processing capabilities for data fitting and rapid processing and analysis.In recent years,it has been applied to microgrid applications.With the development of new energy sources,the consumption of new energy sources under the microgrid has prompted people to study energy optimization within the day.Energy storage devices are a key part of it.Accurately predicting its storage capacity is the task of achieving optimal power scheduling in the microgrid.one.In order to realize the real-time capacity online prediction of energy storage devices under the microgrid,a deep confidence network(DBN)structure in deep learning is used to establish a capacity prediction model for energy storage devices.In order to verify the effectiveness of the model,this paper first analyzes the basic principles of DBN.Through the study of various parameters in DBN,on the basis of experiments,the network topology of the model is determined,including the number of input nodes,the number of hidden layers,and The number of layers of the network.At the same time,using BP neural network to integrate with DBN,make full use of the characteristics of BP algorithm back propagation,fine-tune the parameters in the DBN energy storage device capacity recognition model,thus establishing an improved capacity prediction model.Secondly,compared with the traditional method,the obtained data set is compared with the simulation results of BP and PSO-BP model with optimized particle swarm optimization PSO.The results show that the prediction model using PSO-BP neural network is obviously superior Because of the traditional BP neural network recognition model,the improved DBN prediction model has better performance,higher prediction accuracy and faster speed.The experimental results show that DBN is more suitable for the prediction of high-complexity data than traditional methods.The DBN model realizes the real-time capacity online prediction of energy storage devices.After accurately predicting the capacity of energy storage devices,in order to coordinate the optimal scheduling relationship between various energy storage devices,loads and distributed power sources in the microgrid,on the basis of this,continue to deepen the energy of electric vehicles connected to the microgrid Management strategy for optimal control research,combined with the time-of-use electricity price,the charging pile will display the charging and discharging plan by itself,use the "economy" feature to achieve the orderly charging and dischargingpurpose,and find the best electric vehicle charging load the next day based on the predicted load information Time distribution,using economic levers to guide electric vehicles such as "micro-power" to play their role in "shaving the peaks and filling valleys" in the micro-grid,thereby saving considerable operating costs for the operation management.