What if we can tell you, how your component or system is going to fail, before it fails? Does it help you with your troubleshooting? Or does it help you with your repair planning or to scope repairs if you preventively removed a working component?
NLR developed a new tool to diagnose failures using Artificial Intelligence. It is a clever piece of software that looks at the failure modes of previous repairs and the aircraft usage.
We use Artificial Intelligence to determine the relation between aircraft and system usage and the actual failure modes of repaired parts. We can use these relationships to diagnose components or systems and identify the failure modes.
The trust in the results of computerized diagnoses is highly dependent on the transparency of the analyses. To make the outcomes of failure diagnoses acceptable for maintenance personnel, the algorithms use eXplainable Artificial Intelligence.
FD XAI not only identifies the failure modes, it also explains why a specific failure mode occurs (and not another failure mode). The explanation helps maintenance personnel understand the diagnosis, and troubleshoot failures on the line and in the shop.
Failure diagnoses using eXplainable Artificial Intelligence can be performed before a part actually fails. This means that it is a useful tool to determine the failure mode of parts removed in serviceable condition based on predictive indicators. It helps the shops to repair these parts and it reduces no-fault-founds.