摘要
考虑机组组合的电力系统动态经济调度是一个高维复杂的非线性优化问题。提出了一种采用降维思想解决大规模机组组合问题的新方法,降维的方式是将对整个调度周期的优化转化为对每个调度时刻依次、分别优化,即将对矩阵的优化转化为对行向量的优化,降低求解维数。结合离散与连续粒子群(particle swarm optimization,PSO)算法,分别得到当前调度时刻最优的机组组合状态及对应的最优负荷分配。采用初始化策略提高初始解质量,并对机组启停、爬坡等约束条件处理,使寻优都在可行域中进行,结合优先次序法及智能调整策略避免算法早熟。算例表明本文方法在经济性上具有很大的优越性,且可明显减少开机机组数目,对于求解机组数较多的大规模系统更具优势。
Dynamic economic dispatch of power system considering unit commitment is a high dimensional, complex and non-linear optimization problem. In this paper, a new method of dimension reduction was presented to optimize the large scale of unit on/off states and load dispatch problem, which converted from optimizing the entire dispatching cycle to doing each dispatching time respectively and orderly. That is, transform the optimization of matrix to row vector to reduce the solving dimension. Combined with the discrete and continuous particle swarm optimization （PSO） algorithm, the optimal unit combination states and the corresponding optimal units ＇ load dispatch of current dispatching time were obtained successively. Additionally, a new initialization strategy was proposed to improve the quality of the initial solution. Unit constraints, such as unit commitment constraint, lower and upper limits, were handling to make the optimization in the feasible region. The adoption of the priority list and intelligent adjustment strategy contributed to avoiding the premature of algorithm. Finally, the results indicate that this paper＇ s optimization methods have great economic superiority. Moreover, it can significantly decrease the number of generating units contrast with other literature and has more advantages in the large scale system.
出处
《华北电力大学学报（自然科学版）》
CAS
北大核心
2016年第1期32-38,共7页
Journal of North China Electric Power University：Natural Science Edition
基金
中央高校基本科研业务费专项资金资助项目(2015MS128)
国家自然科学基金资助项目(51177043)
河北省自然科学基金资助项目(F2014502081)
关键词
机组组合
降维优化
动态经济调度
大规模系统
双重粒子群算法
unit commitment
dimensionality reduction optimization
dynamic economic dispatch
large scale system
dual particle swarm optimization
