【作者】 陈婵媛；

【导师】 王坚；

【作者基本信息】 东华大学 ， 电力电子与电力传动， 2007， 硕士

【摘要】 电力市场化是21世纪全球电力工业发展的趋势，是近年来电力工业改革的方向。随着电力系统的迅速发展，电网向着大容量，超高压的联合电网发展，电力系统运行的经济性越来越为人们所重视，电力工业将从传统的运行管理机制向着电力市场运行管理机制转变，使全社会从电力市场的机制转变中获得更大的经济效益和社会效益。电力市场的运行目标是：在满足系统安全稳定运行的条件下，促进发电厂的竞争，以发电成本，网损，辅助服务等方面成本之和最低为优化目标，根据机组报价，确定发电计划，实时调度各个发电公司的机组发电，以满足用电负荷要求。总之，电力市场化的根本目的就是通过建立一个充满竞争和选择的电力工业运营环境以提高整个电力工业的经济效益和社会效益。在竞争的电力市场中，进行优化调度所追求目标是多方面的，可靠性与经济性相协调的目标是定位在一定的资源基础上的各方面综合效益的最大化。本文成功地将遗传算法的基本原理和方法应用于电力系统日有功负荷优化调度的运行中。首先，在研究遗传算法(GeneticAlgorithm，GA)的基础上，本文提出一种应用改进的遗传算法来求解电力系统经济负荷分配的新方法，就是在给定的系统运行方式下，在机组之间合理地分配出力，使全电力系统达到最大的经济性(发电成本为最小)，由此来节约能源。同时，给电厂带来巨大的经济效益。本文建立了电力系统经济负荷分配的模型，在约束条件中考虑了传统算法中通常省略或者简化了的网络损耗的条件。文中对遗传算法参数的编码与译码，初始群体的设定，适应度函数的建立，以及遗传操作和关键参数的确定，都进行了详尽的论述，对于启发式的传统算法和应用于电力系统优化调度的其他的一些智能算法进行了比较，并分别从适应度函数值的标定和增加群体多样性两方面着手解决，在提高算法的收敛速度，防止早熟收敛现象上作了极大的改善。最后通过运算实例证明，运用改进的遗传算法能够求解具有复杂约束条件下的非线性优化问题，适合该问题的求解。与传统的迭代法和一般遗传算法相比，该算法编程简洁，搜索速度快，求解精度高，易于掌握，是解决电力系统经济负荷分配问题的有效手段。本文在用遗传算法解决电力系统经济负荷分配问题的基础上，针对电力市场的特点，提出了一种更适合于现实背景的新的日有功负荷多目标优化分配的策略，该策略同时考虑了发电燃料总耗量，有害气体NOx的排放量和电网收益等三个目标，兼顾了市场各方利益和环境保护。文中建立了三目标优化的模型，并将其应用到山西电网高平县五个发电厂的日有功负荷的模拟优化调度中去。三个目标函数通过改进后的层次分析法加权，或者模糊决策的方法，转化成为单目标问题，最后采用改进的遗传算法求解。运算结果表明了该模型的正确性和优越性，该策略对可持续发展战略和电力市场日交易合同的制定具有重要的现实意义。更多还原

【Abstract】 Market-orientation of electric power is the trends of global power industry development and innovation as well. With the rapid development of electric power system and large capacity & high voltage power network, people are more and more concerned about the economic benefits of power system operation. The power industry will shift from traditional management to market management, which can bring more economic and social benefits.The targets of power market operations are: promote competition between power plants, optimize costs, make daily plans, save energy, protect environment etc. on the basis of safe and stable power system operation.In a word, the dependable and economical aim of modern power market optimization is more and more oriented in the utilizing definite resource and creating maximum benefits than anything else. This article applies the basic principals and methods of genetic algorithm (GA) to the operation of optimized daily active power dispatch. First of all, the article proposes a new way of power dispatch with the application of the improved GA, that is, establishing proper allocation among generating sets to achieve minimum costs and maximum benefits. Secondly, the article sets up the model of economic load dispatch (ELD), taking into consideration of network loss. Thirdly, the article has a detailed elaboration of coding & decoding, population initialization, fitness function establishment and determination of genetic operation and key parameter. It also compares different methods of optimized dispatch based on genetic algorithm and heuristic technique, increases computational rate and avoids premature convergence by enacting the fitness function and increasing population variety. Simulation results show that this improved GA can be used to solve the complicated, nonlinear, constrained optimization problems. Comparing with traditional method and common GA, this improved GA is not only fast but also simple, effective and easy to be well applied to solve ELD problem.At the same time, in view of the characteristics of electric power market, the article proposes a new strategy of multi-objective daily active power dispatch, taking into consideration of total fuel consumption, NOx emission, grid profit and environment protection as well. Through weight value to three objective functions by AHP (analytic hierarchy process) or fuzzy objective membership definition, the multi-objective programming problem is reformulated into non-linear single objective programming problem. With improved GA, the correctness and advantage of the model are proved by the results of calculation examples. This model was used into the Shanxi power network, The results show the correct and advantage of this method which is significant for sustainable development in electric power market.更多还原