(424g) Advanced Process Control for Chemical Batch Reactors Applied to a Polymerization Process

Advanced Process Control (?APC') has become widespread and even a commodity tool for certain continuous processes. APC encompasses model based control solutions, where control and on-line optimization go hand in hand, and a Model Predictive Controller is often used for this purpose.

The advantage of this technology is straightforward: while delivering a significantly improved process control result, it enables to optimize your process on-line taking operational constraints into account.

However, until recently these tools could not be applied to chemical reaction batch processes. The reasons for this are twofold: on the one hand the general APC tools as applied in continuous processes cannot deal with the more complex nature of batch processes (changing dynamics, nonlinear responses). On the other hand the modeling effort and related cost would also be prohibitive to implement an APC project on batch processes with a reasonable return on investment. This would even be more true in a multi-product multi-reactor environment.

Therefore a new strategy has been developed for chemical batch processes. A hybrid modeling solution is proposed which needs a limited engineering input, and delivers a high precision representation of the process. The hybrid model contains both a rigorous mechanistic process descriptions as well as a parametric part. The mechanistic part is used to model often encountered devices in a batch process, like half tube jackets, coils, external heat exchangers, condensers and an overall mass and heat balance. The parametric part is used to describe the reaction kinetics.

This non-linear model is then used within the Model Predictive Control solution. This solution contains an EKF based observer, as well as an interbatch control and observer solution.

Objective of the control solution is to deliver improved control performance, however in a lot of applications batch cycle time optimization is the prime objective. As such the reaction phase of the batch is minimized while keeping the process within the allowable constraints (available heat exchange duty, adiabatic temperature, quality aspects?).

The above mentioned APC methodology has now been applied to several industrial processes. An example will be shown on a polymerization process.