Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
696223 | Automatica | 2012 | 7 Pages |
This brief paper investigates the optimal selection of online controllers for a linear system with actuator faults in the framework of H∞H∞ control. The occurrence of actuator faults is modeled as a Markov chain to predict future faults. A compatible relationship of H∞H∞ controllers is proposed for successive fault modes. In order to shrink the set of all H∞H∞ controllers, seven methods are presented to construct a small set of controllers for different demands. Then five selection schemes are proposed to choose a proper controller from a given set of controllers for a given fault mode. Another selection scheme and a pre-selection scheme are presented when there might be false fault detection and identification. Considered together, they provide a powerful solution for all kinds of demands. A motivating example is used to illustrate the proposed methods.