Anticipatory reactive power reserve maximization using differential evolution L.D. Arya a , Pushpendra Singh b , L.S. Titare c,⇑ a Department of Electrical Engg., SGSITS, 23-Park Road, Indore, MP 452 003, India b Department of Electrical Engg., VITS, Satna, MP 452 003, India c Department of Electrical Engg., Govt. Engineering College, Jabalpur, MP 482 010, India article info Article history: Received 26 February 2011 Received in revised form 24 August 2011 Accepted 5 September 2011 Available online 21 November 2011 Keywords: Voltage stability Reactive power reserve Generation participation factor Proximity indicator Differential evolutionary algorithm abstract This paper presents an algorithm for anticipatory control of load bus voltages. The algorithm optimizes a set of reactive power control variables and maximizes reactive reserve available at generating buses. Voltage dependent reactive power limits have been accounted. The optimal settings of reactive power control variables have been obtained for next interval predicted loading condition. These optimized set- tings satisfy the operating inequality constraints in predicted load condition as well as in present base case loading conditions. A population based differential evolution strategy has been used for optimiza- tion. Results obtained have been compared with those obtained using another population based tech- nique known as PSO. Ó 2011 Elsevier Ltd. All rights reserved. 1. Introduction The problem of reactive power optimization has played an important role in optimal operation of power system. Reactive power optimization (RPO) has complex and non-linear characteris- tics with large number of inequality constraints. Conventional opti- mization techniques, such as linear programming and non-linear programming, take in advantages in computing speed and conver- gence with the objective function of continuous, differentiable and single peak value [1]. Yet conventional methods cannot handle the discrete–continuous problem in reactive power optimization. Re- cently, computational intelligence-based techniques have been proposed in the application of reactive power optimization such as genetic algorithm (GA), Tabu search, simulated annealing, parti- cle swarm optimization (PSO) and differential evolution (DE). These are considered practical and powerful solution schemes to obtain the global or quasi-global optimum solution to engineering optimi- zation problems. At times such schemes are termed as heuristic optimization techniques [2]. Differential evolution algorithm can obtain high-quality solutions within short calculation time and have stable convergence performance. Wu et al. [3] proposed opti- mal reactive power dispatch using an adaptive genetic algorithm. Yoshida et al. [4] suggested a modiﬁed PSO to control reactive power ﬂow and alleviating voltage limit violations. Zhang and Liu [5] proposed a modiﬁed PSO algorithm to deal with multi-objective reactive power optimization. Varadarajan and Swarup [6] proposed differential evolution algorithm for optimal reactive power dis- patch. Zhang et al. [7] have presented dynamic multi-group self- adaptive differential evolution algorithm for reactive power opti- mization. The problem was a mixed-integer, non-linear optimiza- tion problem with inequality constraints. Availability of reactive power at sources and network transfer capability are two important aspects, which should be considered while rescheduling of reactive power control variables. Nedwick et al. [8] have presented a reac- tive management program for a practical power system. They have discussed a planning goal of supplying system reactive demand by installation of adequately sized and adequately located capacitor banks which will permit the generating unit near to unity power factor. Vaahedi et al. [9] developed a hierarchical optimization scheme, which optimized a set of control variables such that the solution satisﬁed a speciﬁed voltage stability margin. Menezes et al. [10] introduced a methodology for rescheduling reactive power generation of plants and synchronous condenser for maintaining desired level of stability margin. Dong et al. [11] developed an opti- mized reactive reserve management scheme using Bender’s decom- position technique. Yang et al. [12] presented a technique for reactive power planning based on chance constrained program- ming accounting uncertain factors. Generator outputs and load de- mands modeled as speciﬁed probability distribution. Monte-Carlo simulation along with genetic algorithm has been used for solving the optimization problem. Wu et al. [13] described an OPF based approach for assessing the minimal reactive power support for gen- erators in deregulated power systems. He et al. [14] proposed a method to optimize reactive power ﬂow (ORPF) with respect to 0142-0615/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.ijepes.2011.09.011 ⇑ Corresponding author. E-mail addresses: ldarya@rediffmail.com (L.D. Arya), erpsingh@rediffmail.com (P. Singh), lstitare@yahoo.co.in (L.S. Titare). Electrical Power and Energy Systems 35 (2012) 66–73 Contents lists available at SciVerse ScienceDirect Electrical Power and Energy Systems journal homepage: www.elsevier.com/locate/ijepes