【作者】 秦祯芳；

【导师】 余贻鑫；

【作者基本信息】 天津大学 ， 电力系统及其自动化， 2004， 硕士

【摘要】 掌握电力市场中电能需求随电能价格变动的规律（即需求响应规律）对于合理确定电价与市场开拓是极为重要的。然而，由于当前我国零售电力市场尚未开放，市场参与者的选择权有限，而且传统的电能销售模式对寻求用户需求对价格的敏感度（即用户的电量电价弹性系数）分析的技术和信息支持均不完备，致使电量电价弹性系数的分析具有很大难度。 为解决这一问题，本文首先尝试基于“当前，峰谷分时电价政策的执行，使用户可以通过选择用电时间，从而间接地选择电能价格”这一现实，针对辽宁省当前零售市场中的行业用电峰谷分时电价的统计数据，运用了一元线性回归模型，挖掘出了电量电价弹性系数。又鉴于电量是受多种因素影响的，本文进而对一些典型行业选用多元线性回归方法进行了分析，给出了电价及相关经济因素对售电量的影响关系。多元回归模型与前一模型所得到的电价对售电量影响趋势相近,从而验证了前一模型的可行性。 然而，研究表明，在实际情况下，部分用户某一时刻的用电量不仅与该时刻的电价有关，还受到相邻时刻的电价影响。因此，本文进一步建立了电量电价弹性矩阵，分析了由于用户分类及电价分类不同所引起的弹性矩阵结构的不同，并在此基础上，提出了一种弹性矩阵的简化方法，给出了在我国当前零售端电力市场中，电量电价弹性矩阵的求取过程，阐明了以月为单位时自弹性系数与交叉弹性系数间的互补性。该方法不仅可以更好的反映电量电价的实际关系，而且提出了新的数据处理方法，消除了地区差异和时间周期的影响，从而扩大了可用样本数。实际应用表明，本文提出的电量电价弹性系数分析方法和模型能够反映当前市场中的需求规律，具有一定的实用价值。 小样本问题是在我国当前零售电力市场分析中经常遇到的问题，因此，本文又分析了偏最小二乘回归在小样本多元分析中的优势，及其在零售电力市场中的应用领域。实例给出了其在配电商成本分析模型中的应用。而应用偏最小二乘回归求得的成本分析指标，符合实际规律，可以用于对配电商进行评价和分析。 上述方法，已经作为核心算法在辽宁省电力市场分析与决策支持系统——电量电价关系子系统中得以应用实现。更多还原

【Abstract】 In order to determine the electricity price reasonably and to extend the powermarket, it is important to master the principle that the energy consumption quantityvaries with the price in power market (called Demand Response principle). However,the current retail power marker in our country has not opened and the option of themarket participants is limited. On the other hand, there are not enough technologiesand data needed by the customer’s price sensitivity analysis (called Price Elasticity ofDemand) in traditional energy sell patterns. All these have complicated the electricityprice elasticity analysis. To solve this problem, this paper first bases on the fact that the performance ofTime of Use enables users to choose the price indirectly by choosing their using time.And it digs out the price elasticity by using the simple linear regression method,according to the industry consumptions in the current retail power market of provinceLiaoning. Because the energy consumption quantity is affected by many factors,multiple regression method has been used for analysis of the typical industries. Theresults have reflected the relationship between energy consumption quantity and pricetogether with other factors. In the aspect of the tendency of price influence on energyconsumption quantity, the same results have been got from above two models. Thishas proved the feasibility of the first model. However, deeper study has proved that the energy consumption quantity isaffected not only by current price, but also by adjacent prices. Therefore, this paperpresents the constructing process of price elasticity matrix of demand, and analysesthe difference of elasticity matrix structure based on different user classes and pricesorts. Then it presents a simplified method for the elasticity matrix, and shows thecalculating process in our current retail power market by an example. The result hasshown that the self-elasticity coefficients and the cross-elasticity coefficients arecompensative. This method can not only reflect the relation between energyconsumption quantity and price more reasonably, but also enlarge the examplenumber by presenting a new method for data proceeding to remove the effect of placedifference and time cycle. The actual applications have proved that the modelsprovided in this paper can reflect the Demand Response principle in current retailmarket, and own some practical effects. Considering that the multivariate data analysis with few observations is ofteninvolved in current retail power market, this paper has also analyzed the advantages ofpartial least squares regression in multivariate data analysis with few observations andits application areas in retail power market. And an example is given to show its usein cost analysis for distribution traders. The results have shown that the calculatedindexes of cost analysis are reasonable and can be used to estimate the distributiontraders. Above methods have been used as the most important arithmetic in pricesubsystem of power system analysis and decision-making system for LiaoningProvince Power Company.更多还原