摘要
提出一种应用主成分和函数变换来分析多影响因素条件下售电市场灵敏度的新方法。首先建立了多维的售电事务空间模型来对售电市场进行恰当的描述,然后提出了基于该模型的灵敏度分析方法,这种方法应用主成分分析来重构空间,从而消除了各影响因素之间的耦合关系。并在此基础上运用基于函数变换的多元回归分析得到最终的售电量分布函数,进而通过该函数的偏导数进行各种灵敏度分析。最后通过某地区电力市场实际数据验证了这种方法的有效性和实用性。
Based on principal component and function transformation analysis, a new method is presented for the sensitivity analysis of electrical power marketing. A spatial model of electrical consumption is set up to describe the electrical marketing. A model based algorithm is proposed to analyze the sensitivity. The principal component analysis is introduced to reconstruct the space of power consumption and erase the linking relation among the factors in it. Then the multi-regression grounded on function transformation is applied to get the function of electrical consumable distribution, whose partial derivative is the foundation of the sensitivity analysis. The new algorithm is demonstrated promising by some practical data of electric marketing of a certain area.
出处
《电力系统及其自动化学报》
CSCD
北大核心
2005年第4期44-48,共5页
Proceedings of the CSU-EPSA
关键词
灵敏度分析
电力市场营销
主成分分析
非线性函数变换
sensitivity analysis
electrical marketing
principal component analysis
non-linear function transformation
