摘要
天然气需求量预测结果受诸多因素的影响,由于各因素的动态随机不确定性以及各因素相互影响的程度难以确定,故准确预测的难度较大。为了提高预测精度,基于前人的研究成果,选取天然气需求量预测中出现频次较高的地区生产总值、产业结构、人口发展、城镇化率、消费水平等变量作为主要影响因素,采用灰色相对关联度刻画各因素的重要性,采用逐步回归法精细检验各因素的显著性,将兼具重要性和显著性的因素纳入逐步回归双对数需求函数模型,通过逐次剔除不显著的影响因素,筛选出相对重要的影响因素,从而构建具有“最少”解释变量的“最佳”预测新模型,并以川渝地区为例对其天然气需求量进行了预测。研究结果表明:①城镇化率是影响该区天然气需求量最重要的因素,经济发展次之;②预测2020—2030年川渝地区天然气需求量呈稳步增长的趋势,但受城镇化建设及外部经济环境的影响,增长率均有所降低,到2030年川渝地区天然气需求总量将达约458×10^(8) m^(3);③四川地区应加快推进城镇化建设,以此拉动天然气需求量增长,重庆地区则应科学规划城镇化建设进度,重点调控经济发展与天然气消费之间的关系。结论认为,新模型能够很好地拟合样本数据,预测结果优于传统模型且精度高,可以作为确定天然气需求量的重要参考依据。
The prediction results of natural gas demand are influenced by many factors,and the dynamic random uncertainty and the influence degree between different factors can be hardly determined,so accurate prediction of natural gas demand is of high difficulty.In order to increase the prediction accuracy,based on previous research achievement,this paper selects the variables that frequently occur in the prediction of natural gas demand as the main influential factors,including GDP,industrial structure,population development,urbanization rate and consumption level.The relative gray correlation degree was adopted to characterize the importance of different factors,and the stepwise regression method was used to finely detect the significance of each factor.The factors with both importance and significance were included in the stepwise-regression double logarithmic demand function model.And a new"optimal"forecasting model with"minimum"interpretation variables was constructed by removing the nonsignificant influential factors gradually and selecting the more important influential factors.Finally,it was applied to predict the natural gas demand in Sichuan-Chongqing area.And the following research results were obtained.First,urbanization rate is the most important factor influencing natural gas demand in Sichuan-Chongqing area,and economic development takes the second place.Second,it is predicted that natural gas demand in Sichuan-Chongqing area will increase steadily during 2020 to 2030.Due to the impact of urbanization construction and external economic environment,however,the growth rate will decrease slightly.By 2030,the total natural gas demand in Sichuan-Chongqing area will reach approximately 458×10^(8) m^(3).Third,as for the Sichuan area,urbanization construction shall be sped up so as to stimulate the growth of natural gas demand.As for Chongqing area,it is necessary to scientifically plan the progress of urbanization construction,focusing on the relationship control between economic development and natural gas consumption.In conclusion,the new model can well fit the sample data,whose prediction result is superior to that by the traditional model and prediction accuracy is higher,so it can be used as an important reference for the determination of natural gas demand.
作者
李洪兵
张吉军
LI Hongbing;ZHANG Jijun(School of Economics and Management,Southwest Petroleum University,Chengdu,Sichuan 610500,China)
出处
《天然气工业》
EI
CAS
CSCD
北大核心
2021年第4期167-175,共9页
Natural Gas Industry
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)。
关键词
天然气
需求量预测
影响因素
灰色关联度
逐步回归分析
双对数需求函数
城镇化率
川渝地区
Natural gas
Demand Forecast
Influential factor
Gray correlation degree
Stepwise regression analysis
Double logarithmic demand function
Urbanization rate
Sichuan-Chongqing area