Development of a new objective function for optimization methods in automatic voltage regulator systems

The most important parameter affecting power quality in electrical power systems is system voltage. System voltage has a very critical design in terms of power system. The system voltage in the electric power system must be maintained at a nominal constant voltage level. Automatic voltage regulators (AVR) keep the system voltage constant. In an AVR system, it is very important that the PID controller parameters are correctly determined. Various optimization techniques are used to find the best of these parameters. The objective functions used in these techniques are generally ITSE, ITAE, ISE and IAE. However, the objective function needs to vary according to the system, problem and constraints. Therefore, the need to make optimization more than one purpose arises. Obtaining the best objective function by considering more than one purpose for AVRs in a limited time creates many objective functional optimization problems. In this study, the best objective function was investigated to determine the optimal PID gains in an AVR system. The multi-objective optimization problem which has emerged to meet multiple purposes, was reduced to more than one objective functional optimization problem with the most appropriate scalarization technique determined within the scope of this study and each reduced problem was solved separately simultaneously. As a method of optimization in the study, particle swarm optimization, ant colony optimization, artificial bee colony algorithm and optics inspired optimization commonly known in the literature, were used. The determined polynomial objective function was found to achieve much better results than objective functions widely used in the literature.