We have to remember that when HUMS was 'invented' the designers said it would take 10 years to gather enough data to understand what the HUMS is telling the end-user. That would appear to be an underestimate on two counts. The first is the need to extend HUMS usage across the board to expand the spectrum of data accumulated from all types and all applications and the second is to find the tools capable of providing reliable analysis of such a vast quantity of data.
I suspect that the way ahead is to somehow get closer to the problem. Wouldn't it be great if you could implant vibe-sensors into the gear wheels and read the output directly. I wonder if any of the electronics boffins are working on such ideas.
The problems caused by harmonics in a structure with so many rotating components are horrendous and trying to second guess the failure mode is a kind of self defeating process given that identifying a failure mode means that you have put your finger on a weakness. Best answer = remove the weakness. The failure mode is therefore by definition the one that takes you by surprise.
Better sensors could be the way forward.
G.
