Research On The Improved Gravitational Search Algorithm For Reactive Power Optimization Of Power System

Optimal reactive power dispatch(ORPD)is an effective method to improve the power quality and economical efficiency of power system,and one of the most basic and important ways to research optimal operation of power system.With the extension of power system and the increase of power load,ORPD has also drawn greater attention of researchers.Firstly,the related concepts of ORPD are introduced,and the mathematical modelings of single objective ORPD(SOORPD)and multi-objective ORPD(MOORPD)are established,respectively.Gravitational search algorithm(GSA)is applied for solving ORPD problem.Then,for SOORPD problem,three enhaced GSA are presented as follows: the first one is GSA-CSS method which applies the proposed conditional selection strategy(CSS)to handle the constraints of ORPD;the second one is GSAPSO method which introduces the memory property of particle swarm optimization(PSO)algorithm and uses acceleration mechanism to balance the impact of memory and gravitation for particles;and the last one is named as IGSA-CSS method which based on the abovementioned two improved GSA,uses Opposition-based Learning(OBL)to optimize the initial population,for enhancing the stability of algorithm.Next,for MOORPD problem,two enhanced multi-objective GSA are proposed as follows: one is the MOGSA-CPD algorithm,which applies the proposed Constraint-prior Pareto-dominance(CPD)method to handle the constraints for MOORPD problem,besides,given the difference between the single objective optimization and multi-objective optimization problems,in MOGSA-CPD,the evolutionary mechanism of GSA is adjusted.The other is MOEGSA-CPD algorithm,which is combined with the memory behavior of PSO,and intruduces the chaotic motion to improve the search ability of algorithm based on MOGSA-CPD.At last,the proposed GSA-CSS,GSAPSO,IGSA-CSS and the conventional GSA and PSO algorithms are applied for solving SOORPD problem based on MATLAB software.The optimal objectives are respectively the minimum active power loss and minimum voltage deviation.The simulation systems contain three systems with different sizes as follows: IEEE30 buses,IEEE14 buses and IEEE57 buses test systems.Experimental results reveal that the presented CSS has higher efficiency for solving SOORPD compared with traditional penalty function method;the presented GSAPSO method has better search effect compared with conventional GSA;and the presented IGSA-CSS method has the most obvious superiority.What's more,in this paper,the proposed MOGSA-CPD and MOEGSA-CPD methods are used respectively for solving MOORPD problem in IEEE30 buses test system with the objective function of minimizing the active power loss and voltage deviation simultaneously.Experimental results show that both MOGSA-CPD and MOEGSA-CPD methods can solve MOORPD problem effectively,and MOEGSA-CPD can obtain the Pareto optimal solution with higher quality,well-distributed Pareto fronts and better optimal compromise solution.