Tuning of PID parameters with renewed particle swarm optimization

In present study, PSO (Particle Swarm Optimization) technique which is widely preferred for PID (Proportional, Integral, Derivative) controller optimization in the industrial environment is taken into account. Since the effect of PID controller parameters on the controlled system is measurable, many studies have been performed with optimizing algorithms to find optimum parameters. In traditional PSO studies, it is only focused around the best particle for the optimization. On the other hand, rather than the best particle which is selected from inital particles, there is no emphasis on the possibility of another particle being better. In this study, the traditional PSO technique has been improved and a new particle solution has been introduced in each cycle. Therefore, a solution was sought out of the initial candidate solutions. For this purpose, the particle which is worst-cousing end of the each cycle has been replaced new candidate particle. In addition, considering the physical effects of the PID controller on the system, different strategies have been developed for the new candidate particle selection. Traditional and improved PSO techniques are used for PID controller optimization with different numbers of particles on an electromechanical system model. According to the results, the optimal solutions can be obtained within shorter time and less particles. In this way, a PSO technique with a lower trial cost for PID optimization is proposed for real systems.