Research On Short-term Wind Power Generation And Load Forecasting Is Micro Grid

In recent years,the advantages of micro-grid have become more and more obvious,which makes people pay more and more attention to the development of micro grid.The balance of power generation and electricity consumption in micro grid is an important index to determine whether the micro grid is economical and efficient.The defects such as the intermittent and uncontrollability of wind power will produce very big threat to the safety of power grid stability,electricity load at the same time also has some not controllable,these problems will seriously affect the whole micro power grid safe and reliable and efficient operation.In order to meet the reliable power supply demand of micro grid,reduce the power generation cost,save energy and improve the efficiency of energy utilization,this paper puts forward the forecast of short-term wind power and load.If can accurately predict the wind power and load of electric power,will be conducive to develop key planning and power generation plan,rational allocation of the fan output,saving energy,which is the main research content of this project.In this paper,artificial neural network is used to predict short-term wind speed,wind power and load,and the forecast of wind power is carried out on the premise of completing wind speed prediction.Firstly,the artificial neural network algorithm is used to predict the wind speed.The traditional BP neural network algorithm is used to improve BP neural network algorithm and GA-BP neural network algorithm.Based on the analysis of these three algorithms,it is concluded that GA-BP network algorithm is the best.Secondly,linear and nonlinear forecasting methods are used to predict short-term wind power.The linear method is the prediction of the wind power based on the establishment of the power curve.By using the direct method and the maximum probability method,the curve diagram of wind speed and fan power is established,then the corresponding wind power can be obtained by using the predicted wind speed.Nonlinear method is the method of using artificial neural network to predict the wind power,after get the prediction power need to predict the power data of the RMSE,MAE and the correlation coefficient is calculated,and the results are analyzed.The experimental results show that the nonlinear method is superior to the linear method,and the BP network algorithm after GA optimization is better than the traditional BP network algorithm.Next,load short-term prediction,this project adopts BP algorithm and similar day to improve BP algorithm.For the same area,the change rule of load change rule with day and week,by choosing similar day is beneficial to reduce the amount of calculation and improve the prediction accuracy,using grey correlation coefficient to the similar day selection.At the end of this paper,the design and implementation of wind power and load forecasting are presented.Firstly,the flow chart of wind power and load prediction is given,and the prediction is made according to the principle of the first three chapters.Then Modbus communication protocol is used to transmit the predicted data obtained by Socket communication technology to the monitoring system,and finally the power generation power of the controlled power is displayed in the micro-grid monitoring system.