摘要：AbstractA nonlinear fault-tolerant control strategy relying on quantitative physics-based models for a cryogenic combustion bench operation is proposed in this paper. The aim is to improve the reliability of a cryogenic bench operation in the transients and to allow to converge to a wider range of operating points. The fault detection is performed with residual-based methods. The residual is generated by an unknown input observer with an unscented transformed which also allows to reconstruct the unknown input. Then the goal is to provide a fault-tolerant system reconfiguration mechanism with a control law which compensates for the estimated actuator additive faults to maintain the overall system stability. For that purpose we use a model predictive control method on an equivalent system with the reconstructed unknown input. An error feedback and a fault compensation control law is designed in order to minimize an infinite horizon cost function within the framework of linear matrix inequalities. The model and the estimation part were validated on real data from Mascotte test bench (ONERA/CNES), and the reconfiguration control law was validated in realistic simulations.
关键词：KeywordsActuator fault accommodationcryogenic systemfault-tolerant controlnonlinear observernonlinear controlLMIdiscrete-time systemmodel predictive control
