Enabling process simulators to perform control/ optimization

Computer assisted modeling of chemical processes has made tremendous progress in the last decades. Large effort is dedicated both in industry and academic research to develop high-grade computer models that mirror the behaviour of physical/chemical processes. Today computer models frequently yield insight into the dynamics of processes in advance of their realization or at operating points, that are (due to safety or economic reasons) not realizable in existing plants. But at most times these models are not exploited further, than to describe steady state performance, dependent on operating parameters. As the predictive capacity of computer models of chemical processes improves steadily, also methods for controller design based on plant models will gain increasing importance. For the simulation of chemical processes partial differential equations are the method of choice to gain physical models. There the partial differential equations are reduced to large sets of ordinary differential equations through spatial discretization. Starting from appropriate initial conditions a discrete time mapping of the state variables is calculated for the systems of ordinary differential equations. Using just this information taken from existing codes is a promising new field for controller design and optimization in chemical engineering. Besides for an analysis of nonlinear systems dynamics and controller design, model order reduction/identification tools are necessary. Controller design through process simulation codes is a fairly young research topic, which is located at the interface of the well established disciplines of process control and chemical engineering.