Research On Gas Load Forecasting Based On Grey Neural Network

As the west-east gas,Sichuan gas to east and other countries of major gas transmission project completed and put into operation,gas fields in our country change from a single pattern of city gas output to a pattern of all-round,multi-angle,gas network system,marking the gas industry of our country has entered a new stage of development.Especially for city gas load forecasting,often presents uncertainty,show the in homogeneity and jumping.However,the law of gas load is still can be grasped and forecasted.Accurate analysis of the regularity can help the relevant departments to accurate and reliable prediction of load the cities,which provide reliable guidance for the gas department.Based on Jiangsu province and Nanjing city as an example,this thesis separately the long-term natural gas load forecast and the short-term load forecasting.First in Jiangsu province as an example,through the study of gas load and the connection between the various factors,such as economy,GDP,per capita GDP,the total gas,natural gas user amount,the total amount of liquefied petroleum gas(LPG),liquefied petroleum gas(LPG)user variables,such as based on gray and neural network combination forecast model,through the instance validation can be accurately predicted by the city gas load,and can successfully provide gas supply guidance for the related department.Selection in the short-term load forecasting,the average temperature,maximum temperature,minimum temperature,the types of weather and date as variable,taking Nanjing as an example,based on the gray and neural network combination forecast model,carried on the short-term gas load forecast empirical analysis.Finally,the prediction model of error analysis proved that grey neural network combination model for gas load forecast error is below five percent,is the reasonable model to provide reliable guidance for the related department.