【作者】 李金超；

【导师】 乞建勋； 牛东晓；

【作者基本信息】 华北电力大学（北京） ， 技术经济及管理， 2009， 博士

【摘要】 电力工业作为能源产业的重要分支,是国民经济的基础产业,是经济发展和社会进步的重要保障。而节能型电力运营机制是实现电力工业可持续发展的必然选择。在当前电力体制下,由于电网公司作用于电力工业的发输配售各个环节,其在实现电力工业可持续发展当中发挥着重要作用,因此如何推动电网公司积极发挥主体作用,运用节能发电调度与电力需求侧管理等手段建设节能型电力运营机制成为目前我国迫切需要研究解决的前沿问题。进而研究探讨建立节能发电优化调度方法、制定需求侧管理措施以及构建节能型电力运营能力评估体系与评价模型等内容具有重要的理论价值和实际意义。论文的主要创新点为:1.建立了基于协整与支持向量机的中长期电量预测方法。通过协整技术建立了能源系统中变量间的长期均衡关系用于电量预测,而后通过支持向量机对协整预测误差进一步修正。计算结果表明此方法提高了预测的可靠性与预测精度。在运用此方法对我国进行中长期电量预测的基础上,对中国未来的节能潜力进行了分析预测。2.构建了节能发电调度指标体系,建立了基于熵权与改进多粒子群算法的组合优化调度方法,实现了对发电侧资源的优化调度。从安全、节能、经济角度选取供电成本、供电煤耗、机组可靠性、厂用电率四个指标,而后运用熵权法计算上述指标的权重,进而构建最优调度目标函数,最后运用改进多粒子群算法实现优化调度。计算结果表明,此方法的计算结果优于遗传算法与模拟退火算法,达到了在降低总耗煤量的同时,其他指标不受过多损失的目的。3.从电源供电能力、电网输配电能力、需求侧用电能力三个角度出发建立了节能型电力运营能力评价指标体系。体系的建立有助于评估电力运营的状态,并且有利于推动电网公司建设节能型电力运营机制,提升电力工业的可持续性。4.建立了循序递进组合评价方法。首先,采用单一综合评价方法进行评价,对于不一致的评价结果,通过Spearman等级相关系数验证各单一评价方法间是否具有相关性;对存在相关性的单一评价方法采用组合评价法对其结果进行组合,若所得评价结果仍不一致,采用Spearman等级相关系数、关联度、最大偏差、平均偏差对组合评价方法的组合效果进行评价,最后运用熵权法对上述组合评价进一步组合得到最终评价结果。此方法克服了单一综合评价方法的片面性和不一致性,使所得评价结果更加客观公正。更多还原

【Abstract】 Electric power industry as an important branch of the energy industry has become the foundation of the national economy industries,economic development and an important guarantee for social progress.The energy-saving operating mechanism for electricity power industry is the inevitable choice in order to achieve sustainable development.In the current power system mechanism,power grid companies are still responsible for the improvement of the electric power industry efficiency.How to promote the power grid companies to play the main role in actively using the energy-saving power generation dispatching,demand-side management and so on to build the energy-saving energy operating model has become a forefront problem,which needed to be solved urgently.Therefore,studying on the establishment of energy-saving optimization dispatching methods,the development of demand-side management measures,as well as construction of evaluation index system and evaluation model for energy-saving electric power system operating ability has important theoretical value and practical significance.The primary innovations are as follows:1.The middle-long power load forecasting method based on co-integration and support vector machine is set up.Firstly,the long-term equilibrium relationship of the variables in energy system is established by means of co-integration technology.The co-integration relationship and error correction model are used to forecast.Secondly,the support vector machine is used to predict the forecasting error of the co-integration and to make the further amendment.The forecasting results show that this model improves the reliability and accuracy.Based on the forecasting results the energy-saving potential of China’s future projections is analyzed.2.The index system of energy saving power generation dispatching is set up.The optimization method for power generation dispatching based on entropy and improved poly-particle swarm algorithm is set up to achieve the optimal scheduling of power generation side resources.The index system includes power supply cost,power supply coal consumption,and reliability of power plant and power consumption rate.The entropy is used to calculate the weights of upper four indexes.Then the objective function is set up.At last,the poly-particle swarm optimization algorithm is used to realize the optimal dispatching.The results show that this method is the best among genetic algorithm and simulated annealing algorithm.Meanwhile,the amount of the total coal consumption is reduced without the other indicators for the purpose have too many losses.3.The evaluation index system for power system energy-saving operating ability is set up from power supply ability,power transmission and distribution ability,and power demand-side electricity usage ability.The index system helps to assess the status of power grid companies,and helps to promote the construction of energy-saving mechanism for electricity operators to enhance the sustainability.4.A sequence and progressive combined evaluation method is set up.The thinking of this method as follows:Firstly,using single evaluation methods to evaluate.If the evaluation results aren’t same,the Spearman rank correlation coefficient is used to calculate the correlation among the results of single evaluation methods.Based on the results,the traditional combined evaluation methods are used to combine the results of the single evaluation method.If the results of the upper combined methods are not same. Here the combined method evaluation indexes,such as Spearman rank correlation coefficient,correlation degree,the max deviation,average deviation,are used to evaluate the upper four combined evaluation methods.The entropy method is used to calculate the weights,and then get the final evaluation result.In this paper,using the indexes system and the data of them,the sequential progressive combined evaluation method is testified.This method overcomes the comprehensive evaluation method of a single one-sidedness and inconsistency,so that evaluation results from a more authentic.更多还原