Summer Daily Load Forecasting Based On Artificial Intelligence Of Qinhuangdao Grid

Electrical power is a widely used form of energy,as the "Two Alternatives"(develop clean energyalternatives to fossil fuels,use electrical poweralternatives to other power)move forward constantly,electrical power will play a more important role in the energy security.Electricity travels at the speed of light,before electrical energy in mass storage technology break through the former,the powergeneration,transmisson,transformation,distributionandutilization must be real-time balance.The power utilization is the basis of the other four links to adjust output.Under the current grid scale,the capacity of the power grid is a fixed value,but the power load is a variable value that following time series.When the peak load exceeds the load capacity of the power grid,the power grid dispatch department must take action to limit the using of power.Without load forecasting,such measures are often urgent,and users who are restricted by electricity use are generally not able to make a proper response,often resulting in larger economic losses.Therefore,the accurate prediction of power load peak has great practical significance.Thethesis describes the research background and significance of the forecast of daily power load,and summarizes the various methods of power daily load forecasting.The operation situation,general load curve characteristics and typical classification load curve characteristics of Qinhuangdao power grid are analyzed in detail,and the influence factors of power load are analyzed in detail.Meteorological elements are important factors affecting the power load,and the power grid load of Qinhuangdao in summer is very sensitive to weather.In the summer,the weather factor is ultimately by influencing the body comfort to influence power load,therefore human body comfort index can be used to replace the temperature,humidity and wind speed as the essence of load forecasting.In this thesis,the average load of human body comfort index,date type and prediction days are used as the conditions for selecting similar days,and the improved particle swarm algorithm is used as the selection method of similar days.Artificial neural network is equipped with powerful nonlinear processing power,which is especially suitable for short-term load forecasting.In this thesis,using the improved BP neural network power load forecasting model is established,and the similar day load data,the human body comfort index and forecasting has 5 load as input of forecasting model,the training set has a high enough typicality and accuracy,so as to improve the learning efficiency of artificial neural network,improve the learning accuracy,more efficient and accurate prediction model is established.At the end of the thesis,in power,on the basis of load forecasting,the basis of practical work and work of Qinhuangdao power grid dispatching operation are proposed,and pointing out the guiding significance of the sale of electricity to the sales department.