Joint synthesisof control and fault detection algorithms: Study of P.I. controller influence

The work reported in this paper focuses on an optimal joint synthesis of controller and fault detection modules in the presence of model uncertainties. The presented method takes the interactions between control and diagnosis into account in order to make more efficient the fault detection performances. A quadratic criterion combining both control and diagnosis aims is defined in a worst case situation according to model uncertainties and disturbances. The cases of proportional and proportional integral controllers are examined in order to study the P.I. controller influence on the joint approach. The residual evaluation is carried out using a threshold computed in an uncertain closed-loop framework.