| More and more distributed power sources are connected to the large power grid,which has a great impact on the operation of the power system.In order to improve the economy and reliability of the microgrid and the large power grid when they are connected to the grid.In this paper,we aim to minimize the operating cost of microgrids and the loss rate of energy storage batteries,and study the microgrid energy scheduling strategy considering demand-side management.At the same time,combined with photovoltaic power forecasting to provide basic data for microgrid energy scheduling,and put forward a consideration The micro-grid protection strategy of load switching provides safety for the application of micro-grid optimal dispatch in actual projects.The specific contents of the study are as follows:(1)Aiming at the problem of excessive noise in photovoltaic prediction data or improper adjustment of hyperparameters,which leads to the inaccuracy of random forest regression(RFR)model in predicting photovoltaic power generation power,a particle based on variational mode decomposition(VMD)combined with improvement is proposed Group(IPSO)optimized random forest regression(RFR)photovoltaic prediction model.This method first selects similar days using the gray correlation coefficient method(GRA),and then uses VMD to decompose the similar day power data into a series of relatively stable sub-modes,highlighting the local feature information of photovoltaic power and reducing the instability of the data Then,the IPSO is used to optimize the hyperparameters in the RFR,the optimized IPSO-RFR model is used to predict each component,and finally the prediction results are superimposed and reconstructed.The example proves that the average absolute percentage error of the model on sunny and rainy days is 5.9%and 10.58%respectively,and the prediction accuracy is relatively high.Provide more accurate forecast of photovoltaic power generation for the study of optimal dispatch of energy considering demand-side management.(2)Establish a mathematical model that considers the demand-side energy scheduling,considering the constraints of lithium battery energy storage in the microgrid system,tie line power constraints,and equilibrium state constraints and transferable load transfer constraints.Cost and lithium battery loss coefficient are optimization goals.Through variable load transfer,energy storage and large grid output,the entire microgrid system is optimally scheduled.An improved non-dominated sorting genetic model solving algorithm is proposed,that is,a non-dominated genetic algorithm is used to introduce a normal distribution crossover(NDX)to combine a binary crossover(Simulated Binary Crossover,SBX)to form a hybrid crossover operator,By adaptively adjusting the weights of the two crossover operators to improve the accuracy of the solution of the non-dominated genetic algorithm.Through the simulation of practical engineering examples,it is verified that the energy scheduling considering demand-side management can make the transferable load shift to the moment when the electricity price is lower and the photovoltaic power generation is more,which improves the economics of microgrid operation.(3)When considering the application of the demand-side management energy optimization scheduling scheme to the actual project of the campus microgrid,when the microgrid runs off-grid,the voltage and frequency may fluctuate because the microgrid system is not controlled by the large power grid.If it is too large and exceeds the adjustable range of the energy storage system,it will make the system unstable.Designed to maintain system stability through cutting machine/load cutting.When carrying out load reduction protection,ensure that important loads will not be cut off at random,classify the loads,and determine the power requirements and power supply measures of power units and equipment according to different load levels to ensure the safety and reliability of the power supply system. |
PDF Full Text Request