L.E. Garza-Castañón, R. Morales-Menéndez, and F.J. Cantú Ortiz (Mexico)
Uncertainty, Fault Tolerant Systems, Bayesian Networks
We propose a new framework to design fault-tolerant mechatronic systems under uncertainty. Mechatronic sys tems are complex mechanical systems where electronics and software are highly integrated. A fault-tolerant scheme in such systems is needed to minimize the overall effect when failures occur in single components or subsystems. Fault-tolerance can be achieved by employing fault diag nosis schemes that determine the severity of faults and acti vate remedial actions to prevent that faults develop into crit ical failures. Uncertainty arises due to the limited amount of information and the cascaded effects caused by faults in other elements of the system. Our work is based on a fault diagnosis scheme extended with remedial actions, and in tegrates BN-FMEA Models, Dynamic Hybrid Probabilis tic Models and the Independent Choice Logic. The pro posed framework allows a flexible structure, able to deal with fault scenarios in large systems, with an adequate way to avoid time consuming inferences.
Important Links:
Go Back
