Robust fault reconstruction using sliding mode observers with nonlinear nominal systems.

This paper is concerned with fault reconstruction based on sliding mode observers; its contribution is twofold: on the one hand, it develops a novel sliding mode observer whose nominal system is nonlinear, thus relaxing former limitations on the subject, namely, rank restrictions, linear outputs, and conservative bounds; on the other hand, based on the referred sliding mode observer, a novel robust fault reconstruction scheme is constructed which can solve a variety of problems that former methodologies cannot treat. Nonlinearities are cast as convex expressions, which enables obtaining design conditions in the form of linear matrix inequalities. Examples are provided to compare the proposal with former methodologies, thus highlighting the contributions.