An Industrial Approach for Efficient Modeling and Advanced Control of Chemical Batch Processes

An industrial implementation of a generic NMPC controller with application to a batch process

“…Therefore, for processes requiring nonlinear control, a hybrid modeling approach (Vandecraen et al [2007]) is chosen, where rigorous modeling is chosen for those model parts that are well understood and a suitably parameterized black-box modeling approach is chosen for the remaining parts. For batch processes, all reactor peripherals (condensors, heat exchangers, cooling coils, .…”

“…are modeled rigorously, whereas the reaction kinetics are modeled by means of an application-specifically parameterized black-box model.This hybrid modeling approach (not to be confused with hybrid models having both continuous and discrete states) leads to discrete-time nonlinear state space models, consisting of an interconnection of rigorous, semi-rigorous and black box submodels. For details, we refer to Vandecraen et al [2007]. The main message here is that in order to obtain sufficiently accurate models with limited effort (cfr.…”

Energy saving in industrial processes using modern data acquisition systems

“…One of the most effective technologies improving industrial processes are the following: predictive control model, neuron networks, and soft sensors technologies. Technologically advanced manufacturing companies, which use innovative processes management and monitoring systems, achieve 20-30% lower production costs than those of similar plants where such systems are not used [1].…”