Font Size: a A A

The Medium And Long Term Power Load Forecasting Based On Improved Grey Model

Posted on:2017-12-10Degree:MasterType:Thesis
Country:ChinaCandidate:Z Z ShenFull Text:PDF
GTID:2322330488491627Subject:Power system and its automation
Abstract/Summary:PDF Full Text Request
Medium and long term load forecasting is very important to power system plan,reconstruction,expansion and development,which is related to the operation of power system securely,economicly and orderly.This paper mainly studies the internal relations between GM(1,1)and its extended model.Besides,it also studies the application of data transformation technology in the grey prediction algorithm,and the structure and application of the multi variable grey variable weight combination model.Specific work is as follows:Firstly,this paper introduces the discrete grey forecasting model(DGM(1,1)model),GM(1,1)model and two extended models,also analyzes the defects of GM(1,1)model.Through the ex~amples of A,B and C provinces,the internal relations of these four models are analyzed,and the shortcomings of the GM(1,1)model are analyzed,and the optimization method is proposed and the following two improved grey prediction models are established:The equal dimensional new information of grey linear regression combined model based on data transformation and residual modification: In this paper,a new type of data transformation function is presented in this paper,which is based on the relationship between the data smoothness and the prediction accuracy of the grey algorithm.Through the ex~ample explores the function y=arccot x^a different a values of GM(1,1)model calculation results of the influence,and illustrates the when using y=arccot x^a is used to transform the data,should obey the actual characteristics of the data,select the appropriate value of a.In this paper,the y=arccot(x^a)data transform function is introduced into the gray linear regression combined algorithm,and for grey linear regression model,using the method of residual modification and the equal dimensional new information to build the equal dimensional new information of grey linear regression combined model based on data transformation and residual modification.Multivariate grey variable right combination model based on residual modification and equal dimensional new information: In this paper,we will analyze the influence of different resolution coefficients on the gray correlation degree,and select the optimal influence factors by the method of grey correlation analysis.For GM(1,N)model can not predict problems using MGM(1,N)model to predict the sequence of the relevant factors,and predicted the introduction of GM(1,N)model,to some extent,solve the above problems.For GM(1,N)model can not predict the problem,using MGM(1,N)model to predict the sequence of related factors,and the results of the GM(1,N)model,to a certain extent,to solve the above problems.For GM(1,1),MGM(1,N),GM(1,N)model boundary value selection,parameter estimation,background value structure is unreasonable and the optimal combination model weight fixed unreasonable problem,A multivariate grey variable right combination model based on residual modification and equal dimensional new information is proposed.In this paper,the two algorithms are applied to the medium and long term load forecasting,the results show that the improved grey prediction algorithm greatly improves the prediction accuracy of the original algorithm,and it has a certain practicality.
Keywords/Search Tags:grey forecasting algorithm, data transformation, mid-long term load, weight changeable combination, multi-variable forecasting
PDF Full Text Request
Related items