LAUSR

Active fault diagnosis is one of the latest frontiers in the field of fault diagnosis, which can improve fault diagnosis performance by redesigning the control input for specific faults. Common methods for active fault diagnosis may be aggressive to systems since the designed control input may affect normal operation. In this article, we aim to develop amicable active fault isolation (AFI) strategy based on control allocation (CA) for dynamic systems with redundant inputs, where the designed control input can enhance fault isolation ability and will not affect normal operation. Within the framework of unknown input observers, the relationship between the measurement capability of the system, the spatial distribution of unknown inputs, and the maximum number of isolable faults (IFs) that occur simultaneously is presented. By changing control allocation algorithms actively, the maximum number of IFs can be increased, under the same measurement capability and the spatial distribution of unknown inputs. CA-based AFI schemes subject to insufficient measurement capability are designed, which can both ensure normal operation and improve fault isolation performance. The effectiveness of the developed AFI strategy is demonstrated through simulation results.