Font Size: a A A

The Research On Combination Forecasting Method Of City Gas Daily Load

Posted on:2020-01-21Degree:MasterType:Thesis
Country:ChinaCandidate:Q LiFull Text:PDF
GTID:2392330575473387Subject:Control Science and Engineering
Abstract/Summary:PDF Full Text Request
As one of the most green and efficient high-quality energy sources in the world,natural gas is widely used by people all over the world.With the rapid development of the global economy and the deterioration of human living environment,the demand for natural gas has increased dramatically.Therefore,it is particularly important to predict the urban gas load.This paper mainly predicts the daily load of city gas.As a result of the year-round range,the temperature and population factors vary little within two months.Therefore,the historical data of gas daily load in a certain city is divided into six groups.The generalized dynamic fuzzy neural network(GD-FNN)and random set theory are used to model and predict the gas daily load.The catastrophe theory is used to predict the possible time point of catastrophe,and then the(average)prediction accuracy and time of each prediction method are obtained respectively.Finally,the advantages and disadvantages of each prediction method are analyzed and compared to obtain conclusions.In view of the research on daily load forecasting method of urban gas,this paper mainly engages in the following work,namely:(1)GRNN,grey-GRNN and gradient-GRNN are used to predict the daily gas load,and the average relative error and prediction accuracy are obtained.The average prediction accuracy of the three methods is 83.734%,85.364% and 86.538%,respectively.In the next prediction,the combination forecasting based on the three forecasting data can improve the accuracy of daily gas load forecasting.(2)GD-FNN is used to forecast the daily load of gas.The forecasting results of GRNN,grey-GRNN and gradient-GRNN above are taken as the data base of GD-FNN.The combined forecasting of the three forecasting data improves the forecasting accuracy of daily load of gas,and the average forecasting accuracy is 90.393%.The average prediction time of each group was 7.668 seconds.(3)Next,the random set theory is used to predict the combined daily gas load,and the urban natural gas consumption reflects the randomness to a certain extent.According to the properties of random set theory and Mahalanobis distance,the average prediction accuracy is 92.389%,and the average prediction time of each group is 0.769 seconds.The method of random set theory is compared with GD-FNN method.(4)Finally,the catastrophe theory is used to deal with the catastrophe phenomenon in the prediction process.The cusp-type catastrophe model is established for the influence factors of population and temperature,and the bifurcation point set equation is obtained.The catastrophe manifold equation and the bifurcation set are used to predict the possible future catastrophe.The feasibility of the application of catastrophe model is verified by comparing with the real situation.This paper simulates and calculates on the basis of MATLAB platform to verify the applicability of the theoretical method.
Keywords/Search Tags:Daily Load Forecasting of Urban Gas, Generalized Dynamic Fuzzy Neural Network, Random Set, Catastrophe Theory
PDF Full Text Request
Related items