基于最优组合模型的中国天然气需求量预测  

China’s predicting natural-gas demand based on an optimal combination model

作　　者：何润民[1] 王富平[1,2] 李洪兵 邹晓琴 王莅[1] HE Runmin;WANG Fuping;LI Hongbing;ZOU Xiaoqin;WANG Li(Natural Gas Economic Research Institute,PetroChina Southwest Oil&Gasfield Company,Chengdu,Sichuan 610051,China;Southwest Petroleum University,Chengdu,Sichuan 610500,China)

机构地区：[1]中国石油西南油气田公司天然气经济研究所,四川成都610051 [2]西南石油大学经济管理学院,四川成都610500

出　　处：《天然气技术与经济》2021年第6期50-57,共8页Natural Gas Technology and Economy

基　　金：四川省科技计划项目“四川天然气供需预测预警机制研究”(编号:2021JDR0241)。

摘　　要：为了精准预测天然气需求量,挖掘有效影响因素,避免天然气需求量受繁多动态非线性因素的影响,在分析灰色相对关联度模型的基础上,从GDP、产业结构、人口发展、能源消费总量、能源消费结构、环境规制、单位能耗、收入水平等8个影响因素中,挖掘出6个有效因素,建立了多元线性回归模型和灰色GM(0,N)模型,并以此构建最优组合预测模型,对天然气需求量进行预测。研究结果表明:(1)中国天然气需求量的有效驱动因素包括GDP、人口发展、能源消费总量、单位能耗、能源消费结构、收入水平;(2) GDP发展对天然气需求量增长贡献最大,GDP的增长可促进天然气需求量增加;(3)基于多因素构建的最优组合模型能进一步降低预测误差平方和,具有良好的非线性逼近预测性能;(4)受能源革命、技术进步及经济环境的影响,预测2021—2025年中国天然气需求量增长率呈下降趋势,到2025年中国天然气需求量将达到4 548×10^(8)m^(3)。In order to accurately predict natural-gas demand and determine effective factors to avoid the influence of various dynamic nonlinear factors on this demand, six effective factors were pointed out from eight influencing ones(GDP, industrial structure, population development, total energy consumption, energy consumption structure, environmental regulation, unit energy consumption, and income level) according to an analysis on grey correlation degree model. Then, the multivariate linear regression model and the GM(0,N) model were established. After that, the optimal combination prediction model was also developed for the demand prediction. Results show that(1) the effective driving factors of China’s naturalgas demand include GDP, population development, total energy consumption, unit energy consumption, energy consumption structure, and income level;(2) GDP development makes the greatest contribution to the demand increase, and the GDP increase can further promote the demand increase;(3) the optimal combination model based on multiple factors can greatly reduce the sum of squares of prediction errors and perform well in nonlinear approximation prediction;and(4) it is predicted that, due to the influence of energy revolution, technological progress, and economic environment, the demand growth rate in China will present a decline trend during 2021-2025 and this demand will reach 4548×10^(8)m^(3)in 2025.

关 键 词：天然气 需求量预测 影响因素 GM(0 N)模型 灰色关联度 最优组合 

分 类 号：F426.22[经济管理—产业经济]