Process Performance Models in the Optimal Design of Protein Production Plants

J. PINTO; J. ASENJO; J.M. MONTAGNA; A VECCHIETTI; O.A. IRIBARREN

Breckenridge, Colorado (USA)

Conferencia; 5th. Conference on Foundations on Computer Aided Design (FOCAPD '99); 1999

FOCAPD

In this work we propose a process performance model to simultaneously optimize the design and the operation of multiproduct batch plants. While the constant time and size factors model is the most widespread used to model multiproduct batch processes, the process performance models describe these time and size factors as functions of the process variables selected as optimization variables. Thus the same mathematical model for the plant is used in both approaches, with the process performance models as additional constraints for the second case. The process performance models are obtained from the mass balances and kinetic expressions that describe each unit operation. A plant for producing human Insulin, Vaccine for hepatitis B, Chymosin and a  cryophilic Protease is used as an example. It is shown on that example that the additional degrees of freedom introduced by the process performance models renders a superior design. The contribution of this approach lies in its inherent modularity, which removes the bound on the size of the problems that can be solved. Given a set of values for the process decision variables, the process performance models are able to predict a set of fixed size and time factors. This make a great difference with respect to the global solution approach for which the problem size rapidly becomes bound, thus making the solution impractical.