Self-Organizing Migrating Algorithm used to control a semi-batch chemical reactor

Abstract

Two different approaches are used to control a semi-batch nonlinear reactor dealing with the significant exothermic reaction. This paper compares evolutionary approach represented by the Self - Organizing Migrating Algorithm (SOMA) with PID pole placement controller (PIDPP). The aim of controlled process is to regenerate the largest amount of chromium filter cake (chromium sludge) in the shortest possible time. For safety reasons the temperature inside the reactor should not exceed 100 degrees C. In this study the reactor is controlled by two manipulating inputs: filter cake feeding flow rate and cooling water temperature. The speed of the whole process depends on the filter cake feeding flow rate - in case of a large amount filter cake addition a sharp temperature rise follows. The controller must be able to find optimal strategy between the filter cake dosing and the cooling temperature to reach the shortest possible process time. The SOMA optimization algorithm was created in Mathematica, PIDPP in Matlab/Simulink.