An Electric Oven Identification Using Step Response Based on the Prony Approach

Ovens are devices commonly used to in laboratories and industry provide temperature conditions for drying, cooking and chemical processes. Developing an accurate mathematical model is an important step before designing a controller for this system. In this study, an electric oven was tested with pulse and step inputs. The Prony curve fitting method was used to fit the step response of the oven to that of a typical nth order Linear Time-Invariant (LTI) system. The decision to use the step response for identification was based on two reasons. First, the oven is a high-order dynamic system with input constraints that make it practically impossible to obtain the impulse response. Secondly, the step input ensures the persistence of excitation and is easy to apply. Adapting Prony's method, which usually uses the impulse response, to work with pulse and step responses required some theoretical modifications. Finally, different model orders were evaluated using the Akaike information function, and it was found that a 17th order LTI system best describes the behaviour of the oven system. The discrete and continuous Transfer Functions (TFs) for this well-fitted model were determined.