On the plantwide control of the Acetone-Chloroform-Benzene-Toluene separation process – a two layer control architecture

In industry, the implemention of advanced control methods in distillation processes remained below the 10% threshold, mainly due to the small profit margin that the implementation of these algorithms would have brought compared to standard control algorithms. Given the increasing demand for very high product purity and productivity, increasingly stringent environmental standards and cost reduction requirements, the implementation of advanced control algorithms is an attractive solution for solving these challenges.The paper presents the application of the plantwide methodology, through an in-house modelling and control framework, to a strongly nonlinear process: the separation of an equimolar azeotropic mixture of acetone, chloroform, benzene and toluene (ACBT) with recirculation. The included models are fundamental models, realized through the equations of physico-chemical, energetic, impulse balances, etc., which take into account a set of simplifying hypotheses. The control was approached from a plantwide perspective and was separated on two levels: a fast one, of stabilization, which will control levels and pressures and a slower one, of supervision, based on predictive control algorithms that will set either the references for the lower level or will directly control the slow loops of the process.The results obtained by a unitary approach starting from modeling, estimating the states and applying plantwide control strategy encourage the increasing usage of the strategy industrial distillation processes.