MOVING HORIZON ESTIMATION AND OPTIMAL EXCITATION IN TEMPERATURE OSCILLATION CALORIMETRY

Calorimetry is a frequently used tool for monitoring and control of chemical reactors. In the case of laboratory-scale reactors with small temperature gradients in the jacket, an additional excitation has to be introduced in order to get reliable estimates if both heat of reaction and heat transfer coefficient are estimated. This is usually referred to as temperature oscillation calorimetry (TOC). In this paper, we introduce a moving horizon estimator that evaluates the data obtained by TOC in real time. In contrast to conventional TOC, this approach is not restricted to sinusoidal signals. Thus the question which excitation yields the best estimates is natural here. The standard deviation of the estimate is taken as a reliability measure.