Research And Application Of Short-term Load Forecasting Technology In Micro-grid

The rapid and steady development of ours country’s economy will inevitably be accompanied by the intensificationofconventionalenergyconsumption.The exhaustion of coal,oil and other energy sources becomes more and more obvious.As an independent and controllable power generation and distribution system,microgrid can effectively combine multiple power generation situations in accordance with the local natural environment.While bringing environmental protection and low-emission power supply mode,it overcomes the power shortage problem in the central and western regions of my country.As an important part of the microgrid energy management system,the short-term load forecasting of the microgrid has increasingly become a research focus of researchers.Accurate prediction of the microgrid load can not only guide the orderly operation of the microgrid system,but also provide an important basis for the power supply plan of the large grid.In view of the characteristics of micronetwork load,the short-term load prediction model of BP neural network based on genetic algorithm optimization is used,and the prediction algorithm is applied to the actual case.The main research work is as follows:Firstly,the characteristics of microgrid load and external factors affecting the load are analyzed to provide theoretical basis for the subsequent prediction model.The BP neural network load forecasting model is established according to the characteristics of high robustness and intelligence of BP neural network.In the prediction model,the characteristics of historical load,day type and weather factors are considered,the load data involved are normalized,the day type and meteorological factors are quantified,and the processed data are taken as the sample of the network,and the prediction of the microgrid load is finally realized.Secondly,according to the initial value of BP neural network is unstable and easy to fall into local minimums and other shortcomings,the BP neural network prediction model based on genetic algorithm optimization is proposed.Using the global search ability of genetic algorithm to optimize the BP network,genetic algorithm replaces the BP neural network gradient descent method to solve the BP neural network iteration speed is slow and easy to fall into the local optimal defects,and the genetic algorithm optimized BP network is used for microgrid load forecasting.By comparing the optimized model with the pre-optimized model prediction results,it is shown that the optimized model prediction accuracy is higher.Finally,the energy management platform of microgrid is constructed,and the load prediction of microgrid is embedded into the second stage energy management platform of microgrid.Theload forecasting subsystem includes hardware selection and software implementation,and uses genetic BP network to predict future daily load.Several groups of forecast results are sampled and analyzed to prove the practical value of the short-term load forecasting.After algorithm simulation and application in Xuzhou Xinri micro-grid load,the power supply efficiency has been effectively improved,which has a certain significance for improving the efficiency of power grid operation.There are 26 figures,9 tables and 62 references included in this thesis.