基于混合粒子群算法并计及概率的梯级水电站短期优化调度

针对梯级水电站短期优化调度的不确定性问题,研究了不确定性因素的概率分布规律,并根据实际系统的运行要求,给出了概率分布密度函数的假设检验方法。探索发电用水量与各种随机因素的互动关系及影响机理,构建了一种新的计及概率的梯级水电站短期优化调度策略。把灾变理论、混沌优化思想和基本粒子群算法结合起来,形成一种混合粒子群算法。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,并能从理论上证明其依概率收敛至全局最优解。将混合粒子群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,求解方法简单有效。仿真结果表明,该策略能较好地处理不确定性条件下梯级水电站的短期优化调度问题。

Short-Term Scheduling Optimization of Cascade Hydro Plants Based on Hybrid Particle Swarm Optimization With Probabilistic Analysis

Taking into account the features of uncertainties presented in short-term scheduling optimization framework of cascade hydro plants, the distributions of uncertainties are formulated in this paper, and a hypothesis test is proposed for the density function of real system. Based on the correlations between the discharges and other statistical variables, a novel strategy for short-term scheduling optimization of cascade hydro plants with risk management is presented. A hybrid particle swarm optimization (HPSO), in which the catastrophe theory and chaos optimization is embedded into original particle swarm optimization (PSO), is also presented for the improvement of community diversity and searching space of PSO, and the global convergence property of HPSO is proved in theory. The presented model is simply solved by an effective combination of HPSO and Monte Carlo simulation. The results show that the method is feasible for the short-term scheduling optimization of cascade hydro plants assuming that some parameters are uncertainties.