Research On Load Forecasting And Energy Storage Dispatching Strategy Of Microgrids

In order to cope with the problems of energy shortage and environmental pollution,microgrids that use renewable energy to generate electricity have been rapidly developed in recent years.In order to promote the nearby consumption of renewable energy,multiple microgrids in a certain area are connected to form a multi-microgrid system,which is beneficial to improve the stability of the microgrid system.However,due to the fluctuation of local loads,microgrid systems have problems such as energy management and optimal operation.Therefore,this paper studies the load forecasting and energy storage scheduling of microgrid,quantifies its volatility through load forecasting,and improves the consumption rate of renewable energy by optimizing the scheduling of the microgrid system.The specific research contents are as follows:(1)An integrated load forecasting model based on Convolutional Neural Networks(CNN)and bidirectional Long-Short Term Memory(LSTM)network is proposed.Aiming at the problems of existing methods that are difficult to effectively deal with multivariate time series forecasting and low forecasting accuracy,this paper establishes a load forecasting model based on CNN,bidirectional LSTM and Adaboost algorithm.The CNN neural network is used to learn the local change trend of the input time series,and the low-dimensional feature sequence is extracted;the bidirectional LSTM neural network is used to learn the long-term dependence of the feature sequence from the time positive sequence and the time reverse sequence.In order to further improve the generalization of the prediction model,the Adaboost algorithm is used to integrate the prediction model combining CNN and bidirectional LSTM.The validity of the proposed model for load forecasting is verified by simulation experiments and comparative analysis.(2)An energy storage scheduling strategy for single microgrid system based on the improved Sparrow Search Algorithm(SSA)is proposed.Aiming at the problem of energy storage management in the microgrid system,based on the load forecast information,considering renewable energy generation,power interaction under dynamic electricity prices,and system protection and control constraints,a microgrid energy storage scheduling model is constructed with the goal of the lowest comprehensive cost.Aiming at the blindness of population initialization and the problem of easy to fall into local optimum,introduce Tent chaotic map,Levy flight mutation and random walk strategy to improve it.Finally,the improved SSA algorithm is used to optimize the energy storage scheduling of the single microgrid system.And the effectiveness of the improved algorithm and scheduling strategy is demonstrated through simulation experiments.(3)An energy scheduling strategy for multi-microgrids system based on adaptive grid multi-objective sparrow search algorithm is proposed.On the basis of energy storage dispatching of single microgrid,considering the energy interaction between microgrids,an energy dispatching model of multi-microgrids system is constructed.Aiming at the problem that the sparrow search algorithm cannot solve the multi-objective optimization problem,based on the algorithm,the Pareto domination and external archive set mechanisms are introduced to obtain the optimal solution set,and the optimal individual is selected by the adaptive grid algorithm.An adaptive grid multi-objective sparrow search algorithm is proposed.Finally,the adaptive grid multi-objective sparrow search algorithm is used to solve the multi-microgrid system scheduling model,and a set of Pareto optimal solutions is obtained,and the scheduling results are selected with the goal of minimizing the total operating cost of the system.Experiments show that the proposed multi-microgrids system energy dispatch strategy can effectively improve the consumption rate of renewable energy and reduce the impact on the stability of the main grid.