Optimization Analysis of Nonlinear Process Using Genetic Algorithm

Controlling the nonlinear process is a very challenging task in the process plant, whereby it depends on the practitioners' knowledge and skills. This paper aims at devel-oping Gain Scheduling (GS) based controller tunings to obtain the trade-off controller tun-ings for both servo and regulatory control objectives at the Low, Medium and High oper-ating levels supported by optimization analysis. At first, the research obtains First Order plus Dead Time (FOPDT) models of various operating levels from the Gravity Drained function of LOOP-PRO software. The dynamic characteristics of GA are compared with Particle Swarm Optimization (PSO), which showed GA produced more desirable re-sponses and performance indexes. The analysis also compares process responses and per-formance indexes of GA with manually calculated controller tunings. The overall result shows that GA optimization analysis produces the most reasonable controller tunings for consistent control performance compared to other methods. Ultimately, GA algorithms were adopted into a Graphical User Interface (GUI) of MATLAB software, allowing the automated generation of the controller tunings for the identified models.