Proposal for a semibatch reactor model refinement suitable for its optimal geometry searching

This paper presents results obtained by simulations of the exothermic semi-batch chemical reactor. The task was to suite chemical reactor dimensions to a specific process in such way, so that the process could run most effectively. An objective function which included also the reactor mathematical model was defined. The objective function was than modified to find even better results and necessary limitations and also penalizations were used to improve the dimensions searching. An evolutionary approach represented by the Self-Organizing Migrating Algorithm (SOMA) was used to minimize the defined function. Results obtained from simulations with the newly found dimensions were compared with original dimensions ones. The results showed feasibility of the dimensions design for a concrete case, the newly designed reactor embodied better performance. At the end, the possibility of the mathematical model used for the simulation refinement is mentioned.