Dear Christoph,

Stability in most batch process is strongly related with feasibility of your MPC control actions. For instance, while running a batch  with an exhotermal reaction you aim to have the fastest possible reaction meanwhile you must keep your reaction from delivering more heat than what your heat exchangers can extract to keep the temperature profile for the reaction. If your MPC delivers feasible actions your system will be stable. Do you have any special batch process in mind?

Regards,

Dear Jairo,

thank you for your reply!

I am considering rather simple batch operations regarding freezing and thawing in the food industry. So there is no reaction whatsoever present and the operations themselves are stable (e.g. as dictated by the laws of thermodynamics). However, I made the experience that sometimes the MPC can become infeasible (especially for short horizons). I am considering NMPC for continuous systems.

Best regards Christoph

Dear Christoph,

From my experience even complicated batch reactions with nonlinearities due to the variable response to the energy applied (phase changes and reactions)  can be controlled with good disturbance rejection mechanism over the "nominal" heat exchange process. In some cases good linear regulators can do the job. The point is that most of your so-called nonlinearity is on the disturbance mechanism, rather than on the heating/cooling mechanism. I will keep it simple in that sense.

You are right MPC will be infeasible for short horizon since you demand a task to be completed on a very short time, that is normal. 

To shade some light about the stability, my experience is that the dynamics of batch processes get strongly simplified if you apply a linearisation along the expected trajectory instead of around some operating point.

Best regards, Jairo

Dear Jairo,

thank you for your answer!

Best,

Christoph