Microgrid Energy Management Strategy Based On Short-term Load Forecasting Of Electric Vehicle Charing Stations

With the increasingly serious problems of traditional energy shortages and environmental degradation,energy conservation and emission reduction,full development and utilization of new energy has become the key to solve the problem.As a result,electric vehicles and microgrids have been vigorously developed and applied.As the scale of electric vehicles continues to grow,connecting them to the microgrid can relieve the pressure of the distribution network.However,the randomness of the charging load of electric vehicles will increase the difficulty of energy optimization of the microgrid.First,this paper analyzes the load characteristics of electric vehicle charging stations,and predicts the daily load in advance through the accurate prediction model.Then,the energy optimization model of the microgrid with electric vehicles is established,and the optimal scheduling is carried out according to the predicted value of daily charging load.The economy and reliability of the operation of the microgrid are taken into account.(1)Aiming at the load characteristics and influencing factors of electric vehicle charging stations,a short-term charging station load prediction method based on data mining technology and convolutional neural network model is proposed.First,the improved data mining model was used to conduct cluster analysis on the historical data,and the training samples with similar characteristics were obtained with the forecast date.Then,based on the analysis of the influencing factors of the charging station load,and considering the internal relationship between the main load influencing factors and the historical load,a convolutional neural network prediction model is established;Finally,aiming at the weak generalization ability caused by random initialization of convolutional neural network model parameters,a multi-objective optimization model was established to optimize the network parameters.(2)Aiming at the energy optimization problem of electric vehicles connected to the microgrid,an economic dispatching strategy of the microgrid with electric vehicles was developed based on V2 G service in the isolated island mode.which took into account the controllability and energy storage characteristics of electric vehicles.First,the daily charging load of electric vehicles is predicted in advance for energy optimization.And then,from the perspectives of the microgrid and electric vehicle users,an optimization model with the minimum load fluctuation of the microgrid,the lowest operating cost and the highest revenue of electric vehicle users as the objective function was established.Finally,NSGA-Ⅱ algorithm is used to solve the model.