【作者】 朱建全；

【导师】 吴杰康；

【作者基本信息】 广西大学 ， 电力系统及其自动化， 2008， 硕士

【摘要】 梯级水电站优化调度具有显著的经济效益和社会效益。随着能源短缺的日益加剧和电力体制改革的不断深化,合理利用水资源,提高水电整体质量和效益,对调整能源结构和实施可持续发展战略具有重要的现实意义。从研究梯级水电站的水能优化利用的思路出发,本文研究了不确定性条件下的梯级水电站短期优化调度和水火电力系统短期优化调度问题,并提出一种混合粒子群算法(HPSO)对问题进行求解,在提高水能资源利用效率和降低电力系统运行成本方面具有一定的实际意义。论文以一定时期内可能实现的总的目标利润最大化为目标,在满足一定的置信水平的前提下满足约束条件,基于机会约束规划构建了一种新的梯级水电站短期优化调度策略。模型全面分析了蓄水量、弃水量、前池水位、放水路水位、发电水头之间的关系,并考虑了电价、入库径流量、机组运行状况等不确定因素对梯级水电站短期优化调度问题的影响。该模型可以根据梯级水电站的实际情况协调风险和利润这两个相互矛盾的指标,实现最优化决策。论文分析了电力市场中日前交易和实时交易的关系,考虑了水火电力系统的水力联系和电力联系,针对各种不确定性因素的影响,基于机会约束规划提出了一种新的水火电力系统短期优化调度的不确定性风险管理模型。该方法兼顾日前交易和在此计划下可能存在的实时交易的费用,提高了水力资源的利用程度,提高了电力系统的运行效益,为实现系统的火电机组和用以实时平衡的功率调整的费用最小化提供了新的研究思路。针对粒子群优化(PSO)算法易陷入局部最优的缺点,论文把灾变理论、混沌优化思想和基本粒子群算法结合起来,形成一种HPSO算法。根据粒子群算法和灾变理论都起源于自然界的生命现象的特点,将灾变理论作为粒子群算法早熟收敛的判据,并根据混沌优化的遍历性、随机性和规律性的特点,利用混沌优化的Logistic方程对陷入早熟收敛后聚集在全局最优粒子周围的粒子位置进行重新的调整。给出了收敛至全局最优解的定义,并从概率测度的角度证明HPSO依概率收敛至全局最优解。该算法扩大了种群的搜索空间,增加了种群的多样性,改善了基本粒子群算法摆脱局部极值点的能力,便于处理复杂约束条件,为求解具有复杂约束条件的非线性规划问题提供了一种简单有效的方法。将HPSO算法嵌入蒙特卡罗随机模拟中对本文提出的模型进行求解。该方法不必将复杂的机会约束规划问题转化为确定性问题再进行求解,比传统的方法更为简单有效。算例的计算结果表明了本文提出方法的有效性。更多还原

【Abstract】 Remarkable economic and social benefit is achieved by the optimal scheduling of cascaded hydroelectric power plants. With the problem of energy shortage and the achievements in recent liberalization of the electricity market, it is necessary to make full use of hydropower, and improve its quality and benefit for adjusting energy structure and carry sustainable development strategy into execution. This paper presents a hybrid particle swarm optimization (HPSO) algorithm, short-term scheduling optimization of cascaded hydro plants, and short-term scheduling optimization of thermal power systems with risk management. The achievement of this paper is capable of providing significant reference value for utilizing cascaded hydroelectric power plants effectively and promoting comprehensive benefit of power system.To maximize the possible total objective profit throughout a time period, a novel strategy for short-term scheduling optimization of cascade hydro plants is presented based on chance-constrained programming in which the constraints are met with a specified probability. The detailed representation of cascade hydro plants, which includes water volume, water inflow, water discharge, forebay elevation, tailrace elevation and effective water head, is studied in the proposed strategy. The uncertainties, such as water inflows, electricity prices and unit status are taken into account as well. The two conflicting targets of profit and risk are coordinated preferably according to the practical system, and the optimization of power output for cascade hydro plants is established by the method developed.A novel model of risk management for short-term scheduling optimization of hydrothermal power system based on chance-constrained programming is presented in this paper. Coordinated relationship between day-ahead trading and real-time trading, operational constraints of hydrothermal power system, and uncertainty in hydrothermal power system are considered in an integrated fashion. This proposal model increases generation benefits of hydro electric power plants, reduces operation costs of thermal power plants, advances comprehensive benefits of power systems, and provides a novel research thought for hydro thermal power systems short-term optimal scheduling problem.To overcome the short coming of easily local optimum of PSO, a hybrid particle swarm optimization (HPSO), in which the catastrophe theory and chaos optimization was embedded into PSO, is also presented in this paper. Since the ideas of PSO and catastrophe theory are inspired by natural concepts, it provides an idea that the stagnation of evolutionary strategies of PSO may be described by catastrophe theory. The logistic map of chaos is used to disjoin the particles, which are close to each other around a local minimum according to the ergodicity, stochastic property and regularity of chaos. The definition of global convergence of stochastic algorithm is defined in the paper, and the algorithm presented is proved to converge to the global optimization solution with probability one using Lebesgue measure. The improvement of community diversity and searching space of PSO is achieved, and the proposal strategy is a simple effective approach to solve nonlinear programming problem with complex constraint conditions. The model presented is solved using a combination of HPSO and Monte Carlo simulation because of the advantages of HPSO such as simple concept, easy implementation and robustness. The results have showed that the combination of HPSO and Monte Carlo simulation for short-term scheduling optimization of cascade hydro plants is versatile and efficient.更多还原