Study On The Analysis Of Influencing Factors And Forecasting Model Of The Maximum Demand For Electricity Consumption Of Large Industrial Consumers

For large industrial consumers whose transformer capacity is greater than or equal to 315 k VA,there are two kinds of charge mode for the basic tariff: respectively,according to the transformer capacity and the maximum demand.The transformer capacity is fixed and known,but the maximum demand as a special power load characteristic is unknown and needs to be predicted based on historical information,which increases the difficulty for consumers to select an economical charge mode and thereby may lead to an increase in the cost of electricity.Therefore,it is necessary to design an intelligent energy management and forecasting system which can get an accurate prediction.In addition,it is particularly important to study an accurate and rapid power load characteristic forecasting technology in order to ensure the stable operation of its production,transportation,distribution,consumption and other links,because the characteristic of electricity is that it can not be stored in large quantities.The power load characteristic is affected by many factors,so it is beneficial to improve the prediction accuracy by taking various influencing factors into comprehensive consideration.This thesis focused on the core issues existed in power load characteristic forecast work of large industrial consumers,such as the characteristics of maximum demand,the selection of main influencing factors,and the construction of high-precision forecasting models.Firstly,the overseas and domestic research status of the influencing factor analysis methods and forecasting methods is elaborated in detail.The existing influencing factor analysis methods,power load forecasting methods and their advantages and disadvantages are introduced.And the influencing factors of power load characteristic are deeply analyzed by using the existing influencing factor analysis methods.Then a combinational analysis method based on PCA-WGRA(Principal Component analysis-Weighted Grey Relational Analysis)is put forward,which not only reduces the complexity of too much influencing factor index,but also takes the influence of time into account.The combinational analysis method can perform weighted processing on each correlation coefficient and obtain the weighted correlation degree between each influencing factor and the power load characteristic,thus determining the main influencing factors.Then,the power load characteristic is made a preliminary prediction using the current universal single-variable model of power load characteristic prediction.And the traditional grey prediction theory is studied on the basis of the analysis of the influencing factors in order to improve prediction effect and prediction accuracy.Aimed at the disadvantages of high raw data requirements and fixed model parameters,an improved exponential smoothing grey forecasting method based on PCA-WGRA is proposed by combining the features of maximum demand prediction.Through MATLAB simulation and data tests,it is proved that the method can correctly analyze the influencing factors and on this basis accurately predict the power load characteristic in real time,which expands the application range of traditional grey forecasting model and makes it no longer limited to data forecasting with strong variation regularity.Finally,on the basis of algorithm research,an intelligent energy management and prediction system based on the C/S architecture is set up.The system consists of influencing factors analysis module,maximum demand prediction module,information query module,server module,back office management module,data management module and other modules,which fully considers the electricity consumption characteristics of large industrial consumers and realizes the prediction of power load characteristic and the determination of main influencing factors that are suitable for large industrial consumers.Through application verification,it is proved that the system can realize energy management functions and the quantitative analysis of influencing factors,thus quickly and accurately predicting the maximum demand as a power load characteristic.