Design of sensor fault diagnosis method for nonlinear systems described by linear polynomial matrices formulation: application to a winding machine
Résumé
In this paper, a sensor model-based fault diagnosis method for a particular class of nonlinear systems is developed. A polynomial matrices representation is considered for modeling the dynamic behavior of a class of nonlinear systems. According to nonlinear representation via a polytopic transformation, the nonlinear faulty system can be considered as a nonlinear system with the presence of additive unknown inputs. Under fault isolation conditions, the main contribution of the paper relies on the use of an accurate observer that performs fault detection and isolation over the whole operating range of the nonlinear system. The effectiveness and performance of the proposed method are illustrated via real tests on a winding machine subject to sensor faults.