Optimum Operation of a Beer Filtration Process

Abstract

This paper deals with the optimum operation of a beer filtration process that uses membranes for this task. Due to fouling, the operation requires cleaning, which damages the membranes, and creates a discontinuous operation. The optimal economic operation can be defined in terms of minimizing the number of chemical cleanings, as well as the use of energy, when processing a certain amount of beer in a given time. The problem is hybrid in nature, due to the discontinuities created by the cleanings. The corresponding optimization problem is formulated in the framework of predictive control but integrating the economic operation as target of the controller and different time scales. Also, instead of using binary variables for representing the discontinuities, the problem employs a sequential approach, embedding them in the dynamic simulation of the process model combined with a control parameterization that allows computing the solution in terms of the continuous variables that represent its degrees of freedom. Results of the optimal operation are presented.