Convex Fault Diagnosis of a Three-Degree-of-Freedom Mechanical Crane.

This paper presents a fault detection and estimation method based on a proportional-integral observer applied to a three-degree-of-freedom mechanical crane. Faults are common in this system and can provoke oscillations that generate a loss of performance and stability. A convex linear parameter varying approach is proposed to stabilize the crane and detect and isolate actuator faults to guarantee the crane's performance. The linear matrix inequalities obtained from candidate Lyapunov functions give sufficient conditions to guarantee the fault estimation method. Finally, numerical simulations are proposed to illustrate the method's performance and applicability. [ABSTRACT FROM AUTHOR]