Open Access

Ruven Weiss, Moritz Diehl, Johannes Reuter

• Analysing observability is an important step in the process of designing state feedback controllers. While for linear systems observability has been widely studied and easy-to-check necessary and sufficient conditions are available, for nonlinear systems, such a general recipe does not exist and different classes of systems require different techniques. In this paper, we analyse observability for an industrial heating process where a stripe- shaped plastic workpiece is moving through a heating zone where it is heated up to a specific temperature by applying hot air to its surface through a nozzle. A modeling approach for this process is briefly presented, yielding a nonlinear Ordinary Differential Equation model. Sensitivity-based observability analysis is used to identify unobservable states and make suggestions for addi- tional sensor locations. In practice, however, it is not possible to place additional sensors, so the available measurements are used to implement a simple open-loop state estimator with offset compensation and numerical and experimental results are presented.