采用改进鲸鱼算法的配电网综合优化

Comprehensive Optimization of Distribution Network Using Improved Whale Optimization Algorithm

为了解决配电网重构与无功优化问题,将线路开关与无功补偿容量同时作为控制变量对配电网进行综合优化。建立以网络损耗、电压改善度、负荷均衡度和开关动作次数为目标的配电网综合优化模型。针对传统鲸鱼算法初始种群分布不均、缺少全局交流、容易陷入局部最优等问题,利用Sobol序列生成分布更均匀的初始鲸群,引入自适应权重调整系数,改进越界处理机制,增加样本多样性的同时产生精英粒子。考虑日负荷、分布式电源出力与电动汽车充电负荷的变化,采用信息熵时段划分法进行日负荷分时段动态优化。算法改进后全局搜索与局部勘探能力更加均衡,更容易跳出局部最优,提高了算法寻优效率。分段动态优化减少了开关动作次数。最后,通过算例验证了所提改进算法及优化策略的有效性。

To solve the problems of distribution network reconstruction and reactive power optimization,the line switches and reactive power compensation capacity are used as control variables to comprehensively optimize the distribution network at the same time.A comprehensive optimization model of distribution network is established with network loss,voltage improvement,load equalization and the number of switching actions as its objectives.Aimed at the problems such as uneven distribution of initial population,lack of global communication,and easy to fall into the local optimality when using the traditional whale optimization algorithm(WOA),the Sobol sequence is used to generate a more evenly distributed initial whale population.The adaptive weight adjustment coefficient is introduced,and the out-of-bounds pullback mechanism is improved to make the algorithm generate elite particles while increasing the diversity of samples.The variations in daily load,output from distributed generation,and charging load of electric vehicles are considered,and the information entropy time division method is used to conduct dynamic optimization of daily load in different periods.After its improvement,the global search and local exploration capabilities of the algorithm are more balanced,which makes it easier to jump out of local optimization and improves its optimization efficiency.As a result,the segmentation dynamic optimization reduces the number of switching actions.Finally,the effectiveness of the improved algorithm and the optimization strategy is verified by an example.