结合专家知识的遗传算法在无功规划优化中的应用

APPLICATION OF EXPERT KNOWLEDGE ADOPTED GENETIC ALGORITHM TO OPTIMIZATION OF REACTIVE POWER PLANNING

在遗传算法中采用专家知识辅助寻优 ,有助于改善寻优方向 ,防止过早收敛。由于无功功率不宜远距离传输 ,因此电力系统中的无功平衡和电压调整具有很强的地区性。依据专家知识对少数被选中的个体动态形成本厂、站的就地无功/电压控制的有效变量集进行人工调整 ,可以改善遗传算法的局部寻优能力。该专家知识包括三个主要方面 :消除电压越限、减少投资和降低网损。为了更符合实际情况 ,建立了母线电压对调节裕度的模糊隶属函数。对某实际系统的计算表明 。

Is helpful to improve the direction of searching optimal solution and prevent the premature convergence of algorithm by use of expert knowledge to Genetic Algorithm (GA) for subsidiary optimization. Because it is not suitable to transmit reactive power for long distance, thus, the reactive power balance and voltage regulation should be performed in limited region. According to expert knowledge the ability of searching partial optimal solution of GA can be improved by artificial regulation of local active control variable\|set,which is formed dynamically by a few selected individuals, for controlling reactive power/voltage of local substation and plant. The expert knowledge mainly includes following aspects: eliminating voltage exceeding specified limits, reducing investment and decreasing network power losses. To accord with practical situation better, the fuzzy membership function of busbar voltage to regulation margin is established. The calculation result of a certain power system shows that the global optimal solution can be found more efficiently by GA combined with expert knowledge.