A hierarchical, enhanced information structure-based fault-tolerant control scheme for industrial interconnected systems with directed acyclic topology

This paper is concentrated on the analysis and design scheme of enhanced information structure-based faulttolerant control for industrial interconnected systems with directed acyclic topology. To this end, the extended state -space expressions of the subsystems are first derived, where the unknown state values of the subsystems are represented by available process input/output data. Then, the direct identification of the key subspace matrices closely related to the extended input/output data models of the subsystems is realized by using LQ decomposition. By embedding the projection matrix as an indicator to measure the interaction strength among the subsystems, an enhanced information structure of the interconnected system is constructed. It is followed by an hierarchical, enhanced information structure-based fault-tolerant control scheme within data -driven realization of the residual generators and the subspace predictive controllers. Finally, a numerical example and a benchmark study are provided to demonstrate the performance of the proposed scheme.