Dynamic Maximun Node Reactive Compensa -tion Based Hybrid Algorithm For Reactive Power Planning In Distribution Networks

The purpose of reactive power planning in distribution network is to determine the best compensation points and capacities, in order to avoid the waste of reactive power. It is an integral part to ensure the voltage quality, loss reduction and the security enhancing. At present, there are many methods which are used for the optimal capacitor placement, such as Classical Optimization Algorithm,Artificial Intelligence and Hybrid Algorithm. Among others, the Genetic Algorithms is used the most popular.When a hybrid algorithm is used,an initial population is generally generated on condition that the number of compensation nodes is known beforehand. Moreover, the determination of upper limit and the initial value for each node's capacitor size has not been reported to the author's knowledge. It is worth pointing that the unreasonable compensation capacities and the harsh constraints of nodes'voltage may increase the calculation burden and affect the solution quality. In addition, a good practical algorithm or software should not make too many requests for users, particularly as determining the total numbers of compensation nodes and the limits of reactive power compensation of each node, which are even difficult for professionals.Based on the features of genetic algorithm and radial distribution structure, a novel hybrid algorithm for reactive power planning optimization in distribution systems is proposed, in which both determinate and stochastic features are organically combined. By adopting a method of dynamic upper limit of node compensation with the consideration of both economy and voltage quality, the problem of determining the upper limits and the total number of candidate nodes beforehand by experience is solved and the resulting initial population could become mostly feasible. For the voltage quality problem, a new strategy of considering the voltage restriction by stages is proposed and the adaptive adjustment mechanism of penalty factors is used in the strategy. Moreover, a simulated annealing, an optimization crossing method and a dynamic disaster mechanism are employed in the proposed hybrid algorithm, and a dynamic threshold verification mechanism of node maximum compensation capacities are also adopted.Based on the proposed algorithm, a computer code is developed using the software platform of CEES (Chongqing Electrical Energy Star, Inc). The validity and effectiveness of the proposed algorithm is demonstrated by applying a computer program developed based on the proposed algorithm to an IEEE 33 test system, an IEEE 69 test system and the 85 nodes'real radial distribution networks.