含风力发电的配电网分时段动态无功优化.docVIP

含风力发电的配电网分时段动态无功优化 凌峰，史静 （江苏省电力公司电力经济技术研究院，南京 210003） 摘要：风电场出力具有随机性，将其接入配电网会引起无功补偿设备的动作频繁。为此，提出一种先对风电场有功输出曲线进行分段，再对各个时段进行整体动态无功优化的方法。针对直接对24h风力发电曲线采用整体动态无功优化会引起维数灾导致算法难以收敛的问题，本文采用基于谱系聚类思想的风电场出力曲线的分段法。针对传统粒子群算法收敛到全局寻优能力较差且易陷入局部最优解的确定，采用云模型对粒子群算法的权值进行动态调整。最后，将上述算法运用于改进的IEEE 33节点系统，仿真结果表明本文提出的1001-1390（201）00-0000-00Reactive Power Optimization in Distribution Network With Wind Power LING Feng, SHI Jing (Jiangsu Electric Power Economic Research Institute, Nanjing 210003, China) Abstract: The access of wind farm causes frequent action of reactive compensation equipment in distribution network. This paper presents a new method, which first segments the wind power curve and then uses overall dynamic reactive power optimization of all periods. To deal with the problem of dimension curse and convergence problem caused by direct overall dynamic reactive power optimization on 24h wind power curve, this paper uses a new segmentation method based on pedigree clustering. To address the problem of traditional particle swarm optimization algorithm which converges to global optimal solution slowly and easily falls into local optimal solution, this paper uses cloud model to dynamically adjust the weights in particle swarm algorithm. At last, the above mentioned algorithm is applied to the improved IEEE 33 distribution network. Simulation results show that the proposed dynamic optimization method reduces the optimization time and decreases the action times of control equipment, therefore extends the working life of the facilities. Key Words: pedigree clustering; sub-period; distribution network; overall reactive power optimization; cloud model 0引言 随着能源危机和环保意识的日益加深，在不久的将来，会有越来越多的风电场作为分布式电源接入配电网。然而，风力发电的随机性、波动性会造成投切电容器和调压变压器的频繁动作。静态无功优化策略已不能适应含风力发电的配电网。动态无功优化计及风电场出力的时序变化，从整体上把握各种无功补偿装置的动作次数[1-3]。 目前对含风电场的配网无功优化研究主要集中在静态的无功优化，动态无功优化问题的文献相对较少。文献[4]采用自然分段法对风力曲线进行分段，把动态无功优化转化为静态无功优化,但该方法的分段结果是由经验得到的。文献[5]提出了先通过静态无功优化确定预动作表实现时空解耦，然后对预动作表进行动态调整的动态无功优化算法，但该算法计算量很大。文献[6]针对一般求解方式复杂而配电网相对简单，提