Dynamic Reactive Power Optimization on the Basis of Family Eugenics Based Evolution Algorithms

TIAN Kang,ZHANG Yao,GUO Li,XIA Cheng-jun

DOI: https://doi.org/10.3321/j.issn:1000-3673.2007.07.002

2007-01-01

Abstract:To improve the convergence speed of reactive power optimization in large-scale power grid, a reactive power optimization algorithm on the basis of family eugenics based evolution is proposed. In this method the orthogonal design technique is led into the course of offspring’s breeding inside a family, thus the individual behavior amelioration can be improved, the prematurity can be avoided and the convergence speed at the late period of evolution can be quickened. In addition, an improved mathematical model for dynamic reactive power optimization and corresponding encoding method are adopted; according to the binary codes corresponding to the numbers of adjustable taps which belong to both transformers and reactive power compensating devices, the length of code string is dynamically assigned; the genome is decided by control variables; taking the binary codes of adjustable taps in control variables of genome as constraint condition, the infeasible solution is modified, thereby the computational efficiency is improved. Results of calculation example show that good effect is achieved while the family eugenics based evolution is applied to reactive power optimization.