基于数据挖掘的电网绩效分析
发布时间：2019-07-06 07:56
【摘要】：绩效分析是电网精益化运营的关键环节,通过选用合适的数据挖掘方法,定量分析电网绩效水平,科学评估电网与负荷需求的协调关系,可以评估电网的规划、投资、改造、运行等业务环节,发现电网发展的薄弱环节以及控制成本、有效利用资金、提升电网性能的最佳途径,能够为电网规划与建设提供有力的理论支撑,进而实现电网与经济发展之间的良性互动。 本文在分析电力企业绩效研究现状,并综合考虑电网运营特点、绩效分析要求、实际数据等问题的基础上,并提出了一种基于DEA的电网绩效分析方法。该方法首先以R聚类技术进行指标精选,以Q聚类技术对不同电网进行分类,并对所选取指标和分类的有效性进行统计检验；然后对同类电网采用DEA绩效分析,并利用松弛变量对非DEA有效的电网进行投影分析,从而得到提升电网绩效的最佳路径,是一种精细化绩效分析方法。最后以中电联提供的“十一五”及2011年统计数据进行省级电网绩效实证分析,验证了上述方法的有效性和实用性。 电网与社会经济的发展联系紧密,电网发展速度不应是一成不变的,定量分析电网与负荷需求的协调性是电网绩效分析的一个方面,并可为电网规划提供依据。本文依托国家电网‘十一五”电网统计调查项目,提出一种以Granger因果检验分析电网与负荷需求协调性的系统方法,研究影响电网发展与负荷需求间的相互影响关系。该方法通过对不满足Granger因果分析平稳性检验的原始数据进行差分处理,使得数据序列满足平稳性要求。并以1993-2011年电网统计数据验证了所提出方法的有效性和实用性。 统计数据质量对电网绩效分析效果的影响,电力统计人员素质又直接影响统计数据质量。本文依托电力统计人员素质调查分析,根据素质评价研究基本理论和电力统计人员知识结构现状,构建了电力行业统计人员素质的评价体系,探索了一种基于德尔菲法改进的层次分析方法,逐层计算素质指标权重,并综合计算统计人员素质综合评价得分,并运用2013年浙江省电力统计人员选调培训测试成绩对电力统计人员素质进行实证分析。
[Abstract]:Performance analysis is the key link of lean operation of power grid. By selecting appropriate data mining method, quantitatively analyzing the performance level of power grid, scientifically evaluating the coordinated relationship between power grid and load demand, we can evaluate the planning, investment, transformation, operation and other business links of power grid, find the weak links of power grid development and the best way to control the cost, make effective use of funds and improve the performance of power grid. It can provide powerful theoretical support for power grid planning and construction, and then realize the benign interaction between power grid and economic development. Based on the analysis of the present situation of power enterprise performance research, and considering the characteristics of power grid operation, performance analysis requirements, actual data and other issues, this paper proposes a power grid performance analysis method based on DEA. In this method, R clustering technology is used to select the index, Q clustering technology is used to classify different power networks, and the validity of the selected index and classification is statistically tested. Then, DEA performance analysis is used to analyze the performance of the same kind of power grid, and relaxation variables are used to project the non-DEA effective power grid, so as to obtain the best way to improve the power grid performance, which is a fine performance analysis method. Finally, the performance of provincial power grid is empirically analyzed with the 11th five-year Plan and 2011 statistical data provided by CLP, and the effectiveness and practicability of the above methods are verified. The power grid is closely related to the development of social economy, and the development speed of power grid should not be immutable. Quantitative analysis of the coordination between power grid and load demand is one aspect of power grid performance analysis, and can provide the basis for power grid planning. Based on the statistical investigation project of the 11th five-year Plan of National Power Grid, this paper proposes a system method to analyze the coordination between power grid and load demand by Granger causality test, and studies the interaction between power grid development and load demand. In this method, the original data which do not meet the stationarity test of Granger causality analysis are processed by differential processing, and the data sequence meets the requirements of stationarity. The effectiveness and practicability of the proposed method are verified by the statistical data of power grid from 1993 to 2011. The influence of statistical data quality on the effect of power grid performance analysis, and the quality of power statisticians directly affects the quality of statistical data. Based on the investigation and analysis of the quality of electric power statisticians, according to the basic theory of quality evaluation and the present situation of knowledge structure of electric power statisticians, this paper constructs the evaluation system of the quality of statisticians in electric power industry, explores an improved analytic hierarchy process based on Delphi method, calculates the weight of quality indexes layer by layer, and comprehensively calculates the comprehensive evaluation score of statisticians' quality. The quality of electric power statisticians in Zhejiang Province is empirically analyzed by using the training results of Zhejiang electric power statisticians in 2013.
【学位授予单位】：浙江大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP311.13;TM711
本文编号：2510888
[Abstract]:Performance analysis is the key link of lean operation of power grid. By selecting appropriate data mining method, quantitatively analyzing the performance level of power grid, scientifically evaluating the coordinated relationship between power grid and load demand, we can evaluate the planning, investment, transformation, operation and other business links of power grid, find the weak links of power grid development and the best way to control the cost, make effective use of funds and improve the performance of power grid. It can provide powerful theoretical support for power grid planning and construction, and then realize the benign interaction between power grid and economic development. Based on the analysis of the present situation of power enterprise performance research, and considering the characteristics of power grid operation, performance analysis requirements, actual data and other issues, this paper proposes a power grid performance analysis method based on DEA. In this method, R clustering technology is used to select the index, Q clustering technology is used to classify different power networks, and the validity of the selected index and classification is statistically tested. Then, DEA performance analysis is used to analyze the performance of the same kind of power grid, and relaxation variables are used to project the non-DEA effective power grid, so as to obtain the best way to improve the power grid performance, which is a fine performance analysis method. Finally, the performance of provincial power grid is empirically analyzed with the 11th five-year Plan and 2011 statistical data provided by CLP, and the effectiveness and practicability of the above methods are verified. The power grid is closely related to the development of social economy, and the development speed of power grid should not be immutable. Quantitative analysis of the coordination between power grid and load demand is one aspect of power grid performance analysis, and can provide the basis for power grid planning. Based on the statistical investigation project of the 11th five-year Plan of National Power Grid, this paper proposes a system method to analyze the coordination between power grid and load demand by Granger causality test, and studies the interaction between power grid development and load demand. In this method, the original data which do not meet the stationarity test of Granger causality analysis are processed by differential processing, and the data sequence meets the requirements of stationarity. The effectiveness and practicability of the proposed method are verified by the statistical data of power grid from 1993 to 2011. The influence of statistical data quality on the effect of power grid performance analysis, and the quality of power statisticians directly affects the quality of statistical data. Based on the investigation and analysis of the quality of electric power statisticians, according to the basic theory of quality evaluation and the present situation of knowledge structure of electric power statisticians, this paper constructs the evaluation system of the quality of statisticians in electric power industry, explores an improved analytic hierarchy process based on Delphi method, calculates the weight of quality indexes layer by layer, and comprehensively calculates the comprehensive evaluation score of statisticians' quality. The quality of electric power statisticians in Zhejiang Province is empirically analyzed by using the training results of Zhejiang electric power statisticians in 2013.
【学位授予单位】：浙江大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP311.13;TM711
【参考文献】
相关期刊论文 前10条
1 楼旭明;窦彩兰;汪贵浦;;基于DEA的中国电力改革绩效相对有效性评价[J];当代财经;2006年04期
2 杨华峰;;电力工业经济增长方式的实证研究[J];电力技术经济;2000年03期
3 何川;杨立兵;李晓刚;周浩;;应用VAR模型分析电力期货与现货价格关系[J];电力系统自动化;2009年04期
4 赵金利;张群华;余贻鑫;贾宏杰;杨锦;;输电网网架结构的谱聚类分析算法[J];电力系统及其自动化学报;2009年04期
5 徐岩;刘金生;张亚刚;王增平;;基于模糊聚类理论的电网故障元件定位[J];电网技术;2010年08期
6 赵渊;张夏菲;谢开贵;;非参数自回归方法在短期电力负荷预测中的应用[J];高电压技术;2011年02期
7 林伯强;电力消费与中国经济增长:基于生产函数的研究[J];管理世界;2003年11期
8 闫庆友;丁子娴;;基于DEA模型的电力行业上市公司的经营绩效分析[J];中国管理信息化;2010年11期
9 于良民;;全面加强学习培训 着力提升领导干部队伍素质[J];国家电网;2011年07期
10 何永贵;张占国;王向阳;;基于层次和模糊综合的电力市场力评价研究[J];华北电力大学学报;2007年04期
相关博士学位论文 前1条
1 唐东明;聚类分析及其应用研究[D];电子科技大学;2010年
，本文编号：2510888
本文链接：https://www.wllwen.com/kejilunwen/dianlilw/2510888.html
教材专著
