Fuzzy identification of reactive distillation for acetic acid recovery from waste water

Abstract

Acetic acid constitutes major component of waste water effluent in petrochemical and fine chemical industries but its recovery poses a serious environmental challenge in the industries. In this work, nonlinear model for multi-input multi-output reactive distillation process was developed for the treatment of waste water containing 30% g/g of acetic acid using fuzzy model technique based on autoregressive with exogenous inputs (ARX) models. It was aimed at obtaining a simplified model which should have essential dynamic behaviour of the process suitable for applications such as control design and dynamic optimization. Data sets used for the model development were obtained from simulated model built in CHEMCAD environment. Reflux flow rate and reboiler heat duty were used as input variables while ethyl acetate top product purity and acetic acid fractional conversion were used as output variables. Structured nonlinear parameter optimization method (SNPOM) was used for parameters estimation to effectively characterize the systems. The percentages of fitness for top product ethyl acetate purity and acetic acid fractional conversion were found to be 85.35% and 82.95%, respectively. The residual analysis carried out on the model showed that the proposed model was able to capture nonlinear dynamic behaviours of the process.