【作者】 张国立；

【导师】 李庚银；

【作者基本信息】 华北电力大学（河北） ， 电力系统及其自动化， 2006， 博士

【摘要】 现货市场的功率分配问题是电力市场的核心问题之一,它关系到市场参与各方的利益,是保证电力市场安全、稳定、有序运行的关键问题。尽管已有大量的研究成果,但是由于这一问题的高度复杂性,从功率分配模型到解法仍有许多问题亟待解决,也是实施电力市场必需要解决的问题。本文在发电市场环境,针对PX+ISO市场运营模式,分析了现货市场的竞争模式,提出了“按统一边际电价结算,按报价进行功率分配”的竞价模式。建立了日前市场购电模型,以及考虑汽轮机阀点效应的经济功率分配模型。为了求解功率分配问题,提出了一种新的变异算子和拟单纯形技术。将拟单纯形技术分别与遗传算法、进化算法结合,得到了混合遗传算法和混合进化算法,并用于求解功率分配问题。随着对环境问题的日益重视,由发电引起的环境问题也越来越受到公众的重视。本文以最小化燃料耗量、排放量以及购电费的三目标功率分配问题模型为例。利用模糊数学理论给出了求解多目标规划问题的模糊加权模型和分明等价模型,并用该方法求解了三目标功率分配问题。模糊加权模型可以灵活地调整权重,以适应电力市场不同情况的需要。考虑到发电成本函数以及有害气体排放量函数中的不确定性,利用模糊数学理论建立了模糊功率分配模型和模糊环境/功率分配模型。模糊功率分配问题属于模糊非线性规划问题,本文提出了利用模糊数的序方法直接优化模糊值目标函数,这种方法可以保留更多的有用信息,为决策者提供更多帮助。本文提出了一种新的权重理想点法,用此方法将模糊环境/功率分配问题转化为单目标模糊规划问题,进而利用直接优化模糊值函数的方法进行优化,得出满意解。在分析了已有现货市场功率分配模型的基础上,指出了日前交易对实时交易有较大影响,建立了综合考虑现货市场中的日前交易和实时交易的统一购电模型(Unified Energy Trading Model,UETM),并证明了UETM优于分开考虑的功率分配模型。另外,本文建立的各种功率分配问题模型都是约束非线性优化问题,应用混合遗传算法、混合进化算法等方法对函数进行优化时必须考虑如何处理约束条件。由于常用的惩罚函数法难以确定惩罚因子的大小,本文利用目标函数值和约束违反程度给出了一种约束处理方法,该方法可用于处理一般约束最优化问题。更多还原

【Abstract】 The load dispatch in spot market is one of the kernel problems of electric power market, which relates to the benefits of each participant in the market, and also the key issue to assure safety, reliability of power system, and order operation of the electric power market. Although a lot of achievements have been obtained, there are yet many problems to be solved urgentlyThe bidding model of the spot market is analyzed, the new bidding rule is proposed for a PX-ISO type electric power market, i.e. market clearing price is determined according to system marginal price, and load is dispatched according to the rule of pay-as-bid price. Build purchase electricity model and economic load dispatch model to solve a non-smooth cost function caused by the steam admission valve openings with a ramp rate constraint. In order to solve these problems, a new mutation operator and quasi-simplex technique are proposed. Combine quasi-simpelx technique with genetic algorithm and evolutionary algorithm respectively, develop hybrid genetic algorithm and hybrid evolutionary algorithm, and apply these methods to solve the load dispatch problems.With increase of consciousness for environment protection, public more and more pay attention to the environment pollution problem caused by generation electricity. A load dispatch model with three objective functions considered emission, purchasing cost and coal wastage is built in this paper. This model is multi-objective optimization problem, absolute optimal solutions are not existed in general, to get satisfying solutions, we give the fuzzy weights model and crisp equivalent model to solve multi-objective programming and apply this method to solve the environmental/load dispatch problem proposed in this paper.Because there are uncertainties in cost the function and emission function, we propose a new fuzzy economic load dispatch model and a fuzzy environmental/load dispatch model by using fuzzy mathematics theory. Fuzzy load dispatch is fuzzy nonlinear programming problem. To solve this problem, we does not use maximal satisfying factor method like usual, but develop a new solution method that compare to different points function values by using fuzzy number ranking method to seek for optimal solutions. This method can reserve more useful information and provide much help to the decision maker than the previous method. Fuzzy environmental/load dispatch is fuzzy multi-objective programming problem, in order to get satisfying solutions, a更多还原