| With the implementation of carbon peaking and carbon neutrality goals,energy efficiency in buildings gains more and more attention and research on the application of "source-grid-load-storage" microgrid systems has increased.How to reduce the impact of renewable energy and load demand uncertainty,and to improve renewable energy consumption rate to optimise the configuration of energy storage system is a pressing problem.The paper mainly studies the microgrid day-ahead load forecasting,optimal dispatching strategy for microgrids considering energy storage and demand response strategy.The main research contents in this paper are as follows:（1）The important basis for optimising the allocation and dispatch of energy systems and achieving local consumption of renewable energy sources such as photovoltaics is the accurate prediction of day-ahead load in microgrid.Therefore,a short-term load prediction model that combines Variational Mode Decomposition（VMD）and Long Short-Term Memory（LSTM）is designed in this paper.The model uses VMD to decompose temperature and humidity data and extracts the mode component with the highest correlation with load as the model input feature vector.Aim at the problem of the LSTM can not correct the hidden state vectors in time,a set of optimized features is generated by inception convolution module to extract the hidden state vectors at different moments of the LSTM output layer,which can compensate the shortcomings of LSTM.（2）The basic structure of the microgrid system is described,and the basic mathematical models of wind power,photovoltaic and energy storage are established to analyse the power output characteristics of each subsystem.The process of energy storage charging and discharging and state of charge are described.The Markov-Monte Carlo method is used to generate the scenario of wind and solar energy,and the K-means clustering method is used to reduce the scenario of wind and solar energy,,and a typical scenario of wind and solar energy is selected as the parameters of the microgrid renewable energy output model.（3）The sparrow search algorithm suffers from the problems of easily falling into local convergence and reduced diversity.The tent chaotic mapping,random flight strategy and sinusoidal formula are used to enhance the diversity during the iteration of the algorithm in this paper.The annealing mechanism is combined with the SSA algorithm to accept non-optimal solutions in the early part of the iteration,which can converge to the global optimum in time.Finally,the improved sparrow search algorithm I-SA-SSA is ompared with other classical algorithms to verify the advantages of the algorithm in terms of convergence speed and accuracy.（4）Connecting the energy storage system to the microgrid is an effective means to deal with the uncertainty of renewable energy,while demand response strategy can optimise demand loads and improve the quality of power system operation.Day-ahead forecast loads and typical scenarios of scenic power generation are used as model input parameters,and an optimal dispatch model for microgrids with energy storage systems based on the generation side is built.After that,the impact of demand response strategy on microgrid operation from the demand side is considered further.Finally,the effectiveness of the optimal dispatching policy obtained by solving the I-SA-SSA algorithm is verified using a microgrid system as an example.Figure [54] table [24] reference [58]... |
PDF Full Text Request