基于改进遗传算法的配电网无功优化

为了提高配电网无功优化的收敛速度，结合配电网根节点电压稳定特性，提出以首端已知电压和计算功率向末端递推，且在回代递推支路末端电压中用第K次电压新值的方法，提高了常规支路电流法的收敛速度，在用遗传算法进行无功优化的过程中，提出了局部搜索替换过程，增加了个体多样性，同时加速了适应值较低的个体的淘汰，提高了每一代个体的平均适应值水平，从而显著提高了常规遗传算法的收敛速度，通过对120节点与53节点的实际算例计算，验证了改进方法能提高电网无功优化的收敛速度和优化效果。

Reactive Power Optimization of the Radial Distribution System Based on the Improved Genetic Algorithm

In order to improve the convergence speed of the distribution system, and combine the constant character of the voltage of the root node, this paper puts forward a method, in which we can deduce the formula from the beginning known voltage and computing power to the terminal. And we use the K th new voltage value when circularly deducing the terminal voltage. This method improves the convergence speed that is always low in routine branch current method. This paper also uses a local searching and replacing method which increases the diversity of the individual, and at the same time accelerates the elimination of the individual which has low fitness value to improve the average fitness value of every generation, and remarkably improve the convergence speed of the GA. Computations of practical examples with 120 nodes and 53 nodes illustrate this method to be correct and efficient.