不同维度数据下的能源经济建模及中国能耗峰值预测研究
发布时间：2019-05-17 04:17
【摘要】：开展能源经济内在规律与趋势研究是制定能源战略的基础性工作。国际经验值得借鉴或警示,但国家、行业、技术等异质性决定了任何单个或少数几个国家能源消费的历史经验都很难单独作为其他国家的参考。本文采用不同维度的数据以刻画这些异质性,通过建立自适应分段计量模型,探索更多国家和部门能源消费轨迹差异及其成因,并形成决策参考。自适应分段计量模型通过对入均GDP(：人均投资和人均消费)的分段回归,得到由多段不同斜率的线段组成的拟合线,在结果得到之前,并不能预期其形状和趋势。这避免了常用的二次、三次函数模型对模型形式的提前主观设定。针对经验研究中通常存在的稳健性问题,本文对各部分研究的分段方法、分段数、控制变量等均进行稳健性检验,并对各部分的峰值预测结果进行相互比较,以检验结果的稳健性。本文数据不仅包括60多个大国1960年以来的面板数据,也包括“60多个国家×50多年时间×5个行业”的三维数据。增加了一个行业维度,考虑到了行业间的关联、扩充了研究中的基础信息量、增强了结论的统计显著性,同时也控制了更多的不可观测因素。主要内容包括：(1)以总量能耗为研究对象,讨论能源消费随着经济发展的历史轨迹,结果表明,国际平均意义上,人均能耗首先随人均GDP增加而增加,在人均GDP达到1.9万美元左右(2005年不变价,购买力平价法,下同),人均能耗开始有下降趋势。考虑到能耗主要是由GDP中的投资驱动的,有必要从需求侧出发,区分投资和消费对能源消费的作用效果,依据国际经验讨论中国未来投资率变化：对能耗的影响。高投资率会导致较高的能耗水平以及较晚的达峰时间。依据“能耗一投资”模型,得到高投资率情景下,我国能耗在2032年达到峰值,峰值量为60亿吨标准油(对应人均GDP为2.9万美元)；在低投资率情景下,在2029年达峰,峰值量为53.5亿吨标准袖(对应人均GDP为2.5万美元)。(2)以公路交通部门为研究对象,分析公路交通能耗的国际历史经验,并据此预测原金砖四国未来公路交通的能源消费。结论表明：在国际平均意义上,随着经济增长,人均公路交通能耗先增加,在达到一个峰值点(人均GDP等于1.9万美元)后,开始有下降趋势。中国公路交通能耗将在2039年达到峰值,峰值能耗为6.1亿吨标油,对应的人均GDP为3.9万美元。(3)以分部门能耗为对象,将国家、行业、时间三个维度的能耗数据集结在一个统一的分析框架内,在考虑产业关联以及多类别不可观测因素的基础上讨论各个行业的能耗历史规律,并据此预测中国各部门能耗。结论表明,在国际一般意义上,总体人均能耗及分部门人均能耗(除其他部门)在经济水平较低阶段均随着人均GDP增长而增加,在达到某一经济水平后逐渐趋缓甚至下降。在低投资率情境下,我国总能源消费量在2032年达到,约为151亿吨标准油(对应人均GDP为2.9万美元)；工业部门在2023年达到峰值,峰值量为12亿吨标准油(1.8万美元)；居民部门在2031年达到峰值,峰值量为7.2亿吨标准油(2.7万美元)；交通部门在2036年达到峰值,峰值量为6亿吨标准油0.5万美元)。本文的预测结果是对“自然”情景的预测。如果今后我国采取约束力更强的节能减排政策,实际用能量将比本文的预测值低。本研究结果也为气候政策综合评估模型提供了一个比较基准。
[Abstract]:The study of the internal law and trend of energy economy is the basic work to develop energy strategy. International experience is worth learning or warning, but the heterogeneity of the country, industry, technology and so on determines that any single or a few of the historical experiences of energy consumption in a few countries are very difficult to serve as a reference for other countries. In this paper, data of different dimensions are used to describe these heterogeneity, and through the establishment of a self-adaptive segmentation model, the difference and the origin of energy consumption in more countries and departments are explored, and the decision-making reference is formed. The adaptive segment measurement model obtains the fitting line composed of segments with different slopes by the segment regression to the average GDP (per capita investment and per capita consumption), and the shape and the trend cannot be expected before the result is obtained. This avoids the usual quadratic and cubic function models in advance subjective setting of the model. In this paper, the robustness of the method, the number of segments, the control variables and the like of each part of the study are tested for robustness, and the peak prediction results of each part are compared with each other to check the robustness of the results. The data not only includes the panel data of more than 60 big countries since 1960, but also the three-dimensional data of the "More than 60 countries have over 50 years of time and five industries". An industry dimension is added, the inter-industry association is taken into account, the basic information amount in the research is expanded, the statistical significance of the conclusion is enhanced, and more non-observable factors are also controlled. The main contents include: (1) The total energy consumption is the research object, and the historical track of energy consumption with the development of the economy is discussed. The results show that the per capita energy consumption is first increased with the increase of the per capita GDP in the international average sense, and the per capita GDP is around US $19,000 (the constant price of 2005). The purchasing power parity (PPP), the same below), the per capita energy consumption begins to decline. Considering that the energy consumption is mainly driven by the investment in GDP, it is necessary to distinguish the effect of investment and consumption on energy consumption from the demand side, and discuss the change of the future investment rate of China according to the international experience: the effect on energy consumption. High investment rates lead to higher levels of energy consumption and later time to peak. According to the "energy consumption one investment" model, under the condition of high investment rate, the energy consumption of China reached the peak in 2032, and the peak value is 6 billion tons of standard oil (the corresponding per capita GDP is US $2.9 million); in the low investment rate scenario, at the peak of 2029, The peak value is 53.5 million tons of standard sleeves (corresponding per capita GDP of US $25,000). (2) The highway traffic department is the research object, and the international historical experience of the highway traffic energy consumption is analyzed, and the energy consumption of the future highway traffic of the former BRIC is predicted accordingly. The conclusion shows that, with the growth of the economy, the energy consumption per capita is increased firstly, and the trend of falling is started after a peak point is reached (per capita GDP is equal to US $1.9 million). China's road traffic energy consumption will reach its peak in 2039, with a peak energy consumption of 6.1 billion tons of standard oil, and the corresponding per capita GDP is $3.9 million. (3) the energy consumption data of the three dimensions of the national, the industry and the time are assembled in a unified analysis framework by using the energy consumption of the sub-sector as an object, and the energy consumption history rule of each industry is discussed based on the industrial association and the multi-category non-observable factors, And the energy consumption of each department in China is predicted accordingly. The conclusion shows that, in the international general sense, the overall per capita energy consumption and the per-capita energy consumption of the sub-sector (other than other departments) increase with the increase of the per capita GDP in the lower stage of the economic level, and gradually decrease or even decrease after reaching a certain economic level. In the context of low investment rate, China's total energy consumption was reached in 2032, about 151 billion tons of standard oil (per capita GDP of $2.9 million); the industrial sector reached its peak in 2023 with a peak of 1.2 billion tons of standard oil ($180 million); and the population sector reached its peak in 2031. The peak amount is 7.2 million tons of standard oil ($2.7 million); the transport sector has reached its peak in 2036 with a peak of 600 million tons of standard oil (US $0.5 million). The prediction of this paper is the prediction of the "natural" scenario. If a more binding energy-saving and emission-reduction policy is adopted in the future, the actual energy will be lower than the predicted value of this paper. The results of this study also provide a baseline for the climate policy integrated assessment model.
【学位授予单位】：北京理工大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：F206
本文编号：2478795
[Abstract]:The study of the internal law and trend of energy economy is the basic work to develop energy strategy. International experience is worth learning or warning, but the heterogeneity of the country, industry, technology and so on determines that any single or a few of the historical experiences of energy consumption in a few countries are very difficult to serve as a reference for other countries. In this paper, data of different dimensions are used to describe these heterogeneity, and through the establishment of a self-adaptive segmentation model, the difference and the origin of energy consumption in more countries and departments are explored, and the decision-making reference is formed. The adaptive segment measurement model obtains the fitting line composed of segments with different slopes by the segment regression to the average GDP (per capita investment and per capita consumption), and the shape and the trend cannot be expected before the result is obtained. This avoids the usual quadratic and cubic function models in advance subjective setting of the model. In this paper, the robustness of the method, the number of segments, the control variables and the like of each part of the study are tested for robustness, and the peak prediction results of each part are compared with each other to check the robustness of the results. The data not only includes the panel data of more than 60 big countries since 1960, but also the three-dimensional data of the "More than 60 countries have over 50 years of time and five industries". An industry dimension is added, the inter-industry association is taken into account, the basic information amount in the research is expanded, the statistical significance of the conclusion is enhanced, and more non-observable factors are also controlled. The main contents include: (1) The total energy consumption is the research object, and the historical track of energy consumption with the development of the economy is discussed. The results show that the per capita energy consumption is first increased with the increase of the per capita GDP in the international average sense, and the per capita GDP is around US $19,000 (the constant price of 2005). The purchasing power parity (PPP), the same below), the per capita energy consumption begins to decline. Considering that the energy consumption is mainly driven by the investment in GDP, it is necessary to distinguish the effect of investment and consumption on energy consumption from the demand side, and discuss the change of the future investment rate of China according to the international experience: the effect on energy consumption. High investment rates lead to higher levels of energy consumption and later time to peak. According to the "energy consumption one investment" model, under the condition of high investment rate, the energy consumption of China reached the peak in 2032, and the peak value is 6 billion tons of standard oil (the corresponding per capita GDP is US $2.9 million); in the low investment rate scenario, at the peak of 2029, The peak value is 53.5 million tons of standard sleeves (corresponding per capita GDP of US $25,000). (2) The highway traffic department is the research object, and the international historical experience of the highway traffic energy consumption is analyzed, and the energy consumption of the future highway traffic of the former BRIC is predicted accordingly. The conclusion shows that, with the growth of the economy, the energy consumption per capita is increased firstly, and the trend of falling is started after a peak point is reached (per capita GDP is equal to US $1.9 million). China's road traffic energy consumption will reach its peak in 2039, with a peak energy consumption of 6.1 billion tons of standard oil, and the corresponding per capita GDP is $3.9 million. (3) the energy consumption data of the three dimensions of the national, the industry and the time are assembled in a unified analysis framework by using the energy consumption of the sub-sector as an object, and the energy consumption history rule of each industry is discussed based on the industrial association and the multi-category non-observable factors, And the energy consumption of each department in China is predicted accordingly. The conclusion shows that, in the international general sense, the overall per capita energy consumption and the per-capita energy consumption of the sub-sector (other than other departments) increase with the increase of the per capita GDP in the lower stage of the economic level, and gradually decrease or even decrease after reaching a certain economic level. In the context of low investment rate, China's total energy consumption was reached in 2032, about 151 billion tons of standard oil (per capita GDP of $2.9 million); the industrial sector reached its peak in 2023 with a peak of 1.2 billion tons of standard oil ($180 million); and the population sector reached its peak in 2031. The peak amount is 7.2 million tons of standard oil ($2.7 million); the transport sector has reached its peak in 2036 with a peak of 600 million tons of standard oil (US $0.5 million). The prediction of this paper is the prediction of the "natural" scenario. If a more binding energy-saving and emission-reduction policy is adopted in the future, the actual energy will be lower than the predicted value of this paper. The results of this study also provide a baseline for the climate policy integrated assessment model.
【学位授予单位】：北京理工大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：F206
【参考文献】
相关期刊论文 前3条
1 杜祥琬;刘晓龙;杨波;王振海;康金城;;中国能源发展空间的国际比较研究[J];中国工程科学;2013年06期
2 廖华;魏一鸣;;七国集团与金砖四国能源结构变化特征及启示[J];中国能源;2010年02期
3 朱永彬;王铮;庞丽;王丽娟;邹秀萍;;基于经济模拟的中国能源消费与碳排放高峰预测[J];地理学报;2009年08期
，本文编号：2478795
本文链接：https://www.wllwen.com/jingjilunwen/jiliangjingjilunwen/2478795.html
