A Constrained Chaotic Flower Pollination Algorithm for Solving ORPD Problem

摘要

Optimum reactive power dispatch (ORPD) is inevitable due to its significant role in the modern power system operation and control planning. ORPD is a highly complex, mixed integer, non-linear and constrained problem with discrete and continuous variables. In this present work, a hybrid flower pollination algorithm (FPA) based on chaotic theory was applied to tackle the ORPD problem. Like other metaheuristics, the original FPA often stagnates in local optima. In order to evade this shortcoming and improve the optimization performance, a sinusoidal chaos map was combined with the standard FPA. Moreover, owing to the various non-linear constraints in ORPD issue, a constraint handling technique named superiority of feasible solutions (SF) is embedded into CFPA. CFPA-SF was applied to ORPD problem considering a medium test system IEEE 57-bus for minimizing the real power transmission losses and voltage deviation as two objective functions. The generator bus voltages, tap setting transformers and shunt reactive power sources are set as control variables. The obtained experimental results and statistical test validate the capability and efficiency of CFPA-SF method in comparison to FPA and those approaches reported in the recent researches in solving ORPD issue.