Prognostics CoE Publications

A hierarchical architecture was discussed for setting up fault adaptive control of a generic overactuated system with embedded nominal controllers. A finite horizon bounded optimization problem was developed to redistribute load among a system’s components, where specifications on maximum risk of failure and maximum degradation of nominal performance are used as constraints. The motivation for using a finite horizon optimization problem is that the cost function allows direct specification of the relative importance of maintaining nominal performance and minimizing risk. The development of this approach for a system with active redundant actuation was examined, and simulation results demonstrated the nature of the optimization problem. Research is underway to address computational issues and parameter selection for the finite horizon optimization problem. Future work will further explore the implementation of the presented architecture on EMA hardware, as well as exploring implementations on new systems, which may not not have active redundancy.