能源消费与经济发展的多尺度分析和反演

【作者】 彭远新；

【导师】 林振山；

【作者基本信息】 南京师范大学 ， 自然地理学， 2011， 博士

【摘要】 对能源消费和经济增长关系的判断是制定和实施能源政策的重要依据,进而关系到宏观经济的平稳运行,因此具有十分重要的现实意义和理论价值。本文利用经验模态分解法和反演建模等方法对不同国家和地区、不同产业类型的经济发展和能源消费关系进行分析,研究结果表明：(1)中国GDP分解后得到的IMF分量蕴含着5.44年的准波动周期,能源消费量分解后得到两个IMF分量,分别蕴含5.10年和11.75年两个准波动周期。通过对比波动周期和方差贡献律,发现二者具有波动性、阶段性和非对称性的特点。对GDP和能源消费量进行反演,通过系数分析发现二者呈非线性关系。(2)对50多年来中国煤炭、石油、天然气、电力消费数据进行分解分析,煤炭消费量分解后的IMF分量显示4.64年和11.75年两个准波动周期；石油消费量分解后的IMF分量显示石油消费存在5.26年的准波动周期；天然气消费量分解后得到三个IMF分量,分别蕴含着3.06年、9.40年、21.50年准波动周期；电力消费量分解后得到一个IMF分量,蕴含着6.53年准波动周期。通过对比GDP与各能源类型的波动周期和方差贡献率,发现它们均具有波动性、阶段性和非对称性的特点。分别对GDP和不同能源类型进行反演,发现GDP与各能源类型也是呈非线性关系。(3)对三次产业GDP和能源消费量进行分解分析,结果显示第一产业GDP蕴含3.47年的准波动周期,第一产业能源消费蕴含2.71年和7.50年两个准波动周期；第二产业GDP蕴含一个4.16年准波动周期,第二产业能源消费蕴含一个7.20年准波动周期；第三产业GDP蕴含一个4.00年准波动周期,第三产业能源消费蕴含一个3.64年准波动周期。通过对比各产业GDP与能源消费量的波动周期和方差贡献率,发现均具有波动性、阶段性和非对称性的特点。分别对各产业GDP与能源消费量进行反演分析,发现各产业GDP与能源消费量都呈非线性关系。(4)运用EMD方法对辽宁省、新疆维吾尔自治区经济发展和能源消费总量及不同能源类型进行多尺度分析,通过对比IMF分量相应的波动周期和方差贡献率,结果显示两省GDP和能源消费总量、各能源类型均具有波动性、非对称性、阶段性特征。分别对GDP和能源消费总量及不同类型能源消费量进行反演分析,发现它们均是非线性关系。(5)通过对美国、日本和英国等发达国家经济发展和能源消费进行EMD分析,发现经济发展的趋势项方差贡献率都比较大,说明发达国家经济都具有不断增长的趋势。能源消费的趋势项方差贡献率则存在较大差异。结果显示发达国家经济发展和能源消费量之间是波动性、阶段性、非对称性的特点。同样对印度和巴西进行分解分析,发现印度和巴西的经济发展与能源消费也具有波动性、阶段性、非对称性的特征。采用反演程序对以美国、日本和英国为代表的发达国家和以印度、巴西为代表的发展中国家经济发展和能源消费量进行反演,结果显示不同发展水平的国家经济发展和能源消费量之间都是非线性的。(6)根据数据的EMD分解图像和方差贡献率进行统计动力建模,提出长时间序列建模、短时间序列建模、指数型增长建模三种预测模型。通过分析发现,如果时间序列后期增长率变化不大,长时间序列建模效果较好。如果后期增长率变化较大,则后两种形式建模较好。如果趋势项方差贡献率相对较低,而且IMF分量波动幅度较大的话,后两种形式建模误差值往往高于5%,模拟精度较低,只能采取长时间序列建模的形式。另外,短时间序列建模、指数型增长建模的模拟结果和预测值基本相似,短期内的预测建立指数型增长模型是相对简洁易行的。本文采用的EMD方法和反演建模对不同区域经济发展和能源消费关系进行分析,不仅丰富了相关的研究方法,而且对指导未来可持续发展有一定的借鉴意义。更多还原

【Abstract】 It is very important to study the relation of energy consumption and economic growth in order to formulate and implement energy policy. And it affects economic development. So the study has very important significance and theoretical value.By means of the Empirical Mode Decomposition (EMD) and Inversion Method, the thesis analyzes the relation of energy consumption and economic growth of different countries and regions and various industrial type. The research suggests:(1)Intrinsic Mode Function (IMF) shows that the fluctuation of China’s GDP has 5.44-year quasi-period. And IMFs of China’s energy consumption indicate that energy consumption have 5.10-year and 11.75-year quasi-period. Contrasted quasi-period and variance contribution rate with each other, it can be found that energy consumption and economic growth have properties of fluctuation, periodical and asymmetry. And it suggests that they are nonlinear after Inversion.(2) The consumption data of China’s coal, oil, natural gas, electricity for 50 years were decomposed, it was shown that the IMFs of coal consumption include two quasi-periodic of 4.64 years and 11.75 years. The IMF of oil consumption shows 5.26 years quasi-periodic. The IMFs of natural gas consumption contain 3.06 years,9.40 years,21.50 years of quasi-periodic, respectively. The IMF of electricity consumption shows 6.53 years quasi-period. Compared the GDP with different types of energy of their cycle time and variance contribution rate, it can be found that energy consumption and economic growth have properties of fluctuation, periodical and asymmetry. And it is suggested that they are nonlinear after Inversion.(3) The GDP and energy consumption of three industries were decomposed by the use of EMD. The results show that the GDP of primary industry contain one cycle of 3.47 years. Energy consumption of primary industry include two cycles of 2.71 years and 7.50 years. The GDP of secondary industry implies a quasi-period of 4.16 years, energy consumption of secondary industry show two cycles of 2.71 years and 7.50 years. The GDP of tertiary industry contains 4.00 years quasi-period, and energy consumption of tertiary industry shows one quasi-period of 3.64 years. Compared the GDP with energy of their cycle time and variance contribution rate, it can be found that energy consumption and economic growth have properties of fluctuation, periodical and asymmetry. And it suggests that they are nonlinear after Inversion. (4) Liaoning Province and Xinjiang Uygur Autonomous Region’s GDP and energy consumption were analyzed by EMD. Compared the GDP with energy of their quasi-period and variance contribution rate of IMFs, it can be seen that energy consumption and economic growth have properties of fluctuation, periodical and asymmetry. And it suggests that they are nonlinear after Inversion.(5) The GDP and energy consumption of developed countries were analyzed by EMD, for example, the United States, Japan and the United Kingdom. The results reveal that the variance contribution rate of economic trend were large, And it shows that the trend of GDP has being grown. But variance contribution rate of energy consumption trend were quite different. The results indicate that the developed countries economic development and energy consumption have properties of fluctuation, periodical and asymmetry.The economy and energy consumption data of India and Brazil also were decomposed by EMD. And it was found that they also have properties of fluctuation, periodical and asymmetry. It also suggests that they are nonlinear of different national economy and energy consumption after Inversion.(6) The statistics and dynamic modeling are modeled according to the images of EMD and variance contribution rate. The models include long time series modeling, short time series modeling, and exponential growth model. The analysis shows that long time series modeling is better if the growth rate changes small in time series late stage. The other two forms of modeling are better if the growth rate changes greatly. If the variance contribution rate of trend is relatively low, and IMF component fluctuate largely, error of two models is often higher than 5%, and the simulation accuracy is low, so the long time series modeling is need. In addition, because the results of simulation and prediction are similar for short time series modeling and exponential growth modeling, it is easy to establish exponential growth model for short-term forecast, and it is operated easily.In this paper, the different regional economy development and energy consumption were analyzed by EMD and Inversion model, it not only enrich the research methods, and it can guide sustainable development in the future.更多还原