Heat exchanger identification by using iterative fuzzy observers
Résumé
The principle of fuzzy observers is first illustrated on a general example: the determination of the two parameters of second order systems using a step response. The set of equations describing the system are presented and it is shown that accurate results are obtained, even for a high level of noise. The heat exchanger model is then introduced. It is based on a spatial division of a counter flow heat exchanger into multiple sections. The governing equations are rewritten as a state space representation. The number of sections needed to get accurate results is determined by comparing estimated values to experimental data. Based on the mean value of the root mean squared errors, it is shown that 80 sections is an appropriate value for this heat exchanger. It is then shown that the iterative fuzzy observers can be used to determine the main parameters of the counter flow heat exchanger, i.e. the convection heat transfer coefficients, when in transient state. The final values of these parameters are <3.5 % apart from the values determined by a time consuming trial and error procedure. Finally a sensitivity study is carried out, showing that a ±1.5 % variation of the actual value of the overall heat transfer coefficient corresponds to a ±0.5 % variation of the estimated overall heat transfer coefficient. This study also shows that the fuzzy observers are equally efficient when the heat exchanger is in steady state.