Speaker : Peter Kuehl Abstract: Introduction to Model Predictive Control Control problems are ubiquitous in modern chemical process engineering, ranging from low-level flow control to high-level overall plant control. The vast majority of controllers used for these purposes are proportional-integral (PI) controllers as introduced in the 50s. It can be easily shown however, that these controllers may fail when the processes become nonlinear. Demanding economic needs and environmental regulations ask for new controller concepts that can handle nonlinearities and allow for control in an optimal fashion. One such control scheme is called (Nonlinear) Model Predictive Control (NMPC). The talk aims at pointing out needs and difficulties of process control. After briefly reviewing the concepts of PI control, it introduces the main ideas and characteristics of NMPC. The talk is designed for an audience with a strong scientific background but does not require particular knowledge in the area of process control.