MINLP Optimization of Side-Reactor Column Configuration Based upon Improved Bat Algorithm

The side-reactor column (SRC) configuration is an effective integrated reaction separation technique, but its complex nature such as many nonlinear equations and integer type variables raises difficulty for process optimization. Here, a mixed-integer nonlinear programming (MINLP) approach coupled with simulation software (Aspen Plus) was proposed to design an SRC configuration. The MINLP model was developed to optimize the structural/operating variables simultaneously. The use of Aspen Plus provided high convergence and accuracy of rigorous process model simulation. We improved stochastic algorithm bat algorithm to solve the MINLP model. The model and algorithm were both implemented in the programming software MATLAB. The capability of this optimization approach was manifested by two case studies of methyl acetate and methyl laurate production systems.