2013 AIChE Annual Meeting

(443b) Model Predictive Control of a Large-Scale Process Network Used in the Production of Vinyl Acetate

Authors

TungSheng Tu - Presenter, University of California, Los Angeles, Los Angeles, CA
Matthew Ellis, University of California, Los Angeles
Panagiotis Christofides, University of California, Los Angeles



Vinyl acetate is an important chemical for its use in the production of polyvinyl acetate which is subsequently used in manufacturing polyvinyl acetate and other vinyl acetate co-polymers. Polyvinyl acetate is used in the production of adhesives and the production of other polymers. Therefore, optimally producing vinyl acetate is of paramount importance to the chemical process industry. To produce vinyl acetate, two raw materials, ethylene and acetic acid, are required as well as a multiple unit process network. Furthermore, the dynamics of the large-scale process network are inherently nonlinear because of nonlinear thermodynamic relationships, rate expressions, multiple recycle streams, and process constraints [1]-[2]. Typically, the control system of such process consists of multiple independent single-input single-output proportional integral derivative (SISO PID) control loops [1]. However, little work has done on studying the application of advanced control systems that can account for the nonlinearities such as model predictive control (MPC) [3]-[4].

Motivated by the lack of applications of MPC to a large-scale process network used in the production of vinyl acetate, we focus on the development and application of two Lyapunov-based model predictive control (LMPC) [4] schemes to a large-scale nonlinear chemical process network used in the production of vinyl acetate. The nonlinear dynamic model of the process consists of 179 state variables and 13 control (manipulated) inputs, and the major processing units of the network include a vaporizer, plug-flow reactor, absorber, process-to-process heat exchanger, separator, and absorber. The two control schemes considered are an LMPC scheme which is formulated with a quadratic cost function and a Lyapunov-based economic model predictive control (LEMPC) scheme [5] which is formulated with an economic (non-quadratic) cost measure. The economic cost measure for the entire process network accounts for the reaction selectivity and the product separation quality. Simulations are carried out to study the economic performance of the closed-loop system under LMPC and under LEMPC formulated with the proposed economic measure. A thorough analysis of the two control schemes will be provided.

References:

  1. Luyben ML, Tyres BD. An industrial design/control study for the vinyl acetate monomer process. Computers & Chemical Engineering. 1998;22:867-877.
  2. Luyben WL. Design and control of a modified vinyl acetate monomer process. Industrial & Engineering Chemistry Research, 2011;50:10136-10147.
  3. Mayne DQ, Rawlings JB, Rao CV, Scokaert POM. Constrained model predictive control: stability and optimality. Automatica. 2000;36:789-814.
  4. Christofides PD, Liu J, Munoz de la Pena D, Networked and Distributed Predictive Control: Methods and Nonlinear Process Network Applications. Advances in Industrial Control Series. London, England: Springer-Verlag, 2011.
  5. Heidarinejad M, Liu J and Christofides PD. Algorithms for improved fixed-time performance of Lyapunov-based economic MPC of nonlinear systems. Journal of Process Control. 2013;23:404-414.