Tek-Tips is the largest IT community on the Internet today!
Members share and learn making Tek-Tips Forums the best source of peer-reviewed technical information on the Internet!
-
Congratulations TugboatEng on being selected by the Eng-Tips community for having the most helpful posts in the forums last week. Way to Go!
Fatigue Test Of UAV Wing
- Thread starter lLouie
- Start date
Ok, so you've done a fatigue test to 20000 cycles, presumably cycles of zero to some load ... limit load ? some %age of limit load ?
First you need to decide on a "safe life factor". For full scale fatigue tests OEMs will frequently use 2 or 3, but you can use more or less ... I mean people aren't going to die if you get it wrong (not unless they are very unlucky), but then maybe there are expensive things on your UAV ?
Then you need to decide on the cycle-to-flight relationship. A very simplistic (and usually conservative) approach is to say 1 limit load cycle represents a flight. A more analytical approach is found in ESDU 69023 (which I think can be found online). But you should consider that this data wasn't collected with UAVs in mind ... I suspect your UAV flies at very low altitude (compared to a large commercial transport). It may be reasonable to assume it's valid, then go collect your own data, with an test article instrumented in an "nz" accelerometer, a "loads monitoring" system.
But my experience is with metallic structures, and you've mentioned your wing is composite. I know composite primary structures is different (and no doubt others will post on it). I understand the most important thing is "BVID" ... possibly Barely Visually Inspectable Damage. I think you need to do a tap test or some other inspection of the test specimen (and use your FEA to determine the most likely places for a disbond) to determine if damage has occurred. I would rather continue the test until you get a failure, but you can always assume failure has occurred and you're back to the two previous points. But a single load test may be less valid for composites, I've heard that they tend to be more sensitive to low loads (high cycle frequency).
Another thing about composites is vulnerability to UV light, generally environmental effects
"Hoffen wir mal, dass alles gut geht !"
General Paulus, Nov 1942, outside Stalingrad after the launch of Operation Uranus.
First you need to decide on a "safe life factor". For full scale fatigue tests OEMs will frequently use 2 or 3, but you can use more or less ... I mean people aren't going to die if you get it wrong (not unless they are very unlucky), but then maybe there are expensive things on your UAV ?
Then you need to decide on the cycle-to-flight relationship. A very simplistic (and usually conservative) approach is to say 1 limit load cycle represents a flight. A more analytical approach is found in ESDU 69023 (which I think can be found online). But you should consider that this data wasn't collected with UAVs in mind ... I suspect your UAV flies at very low altitude (compared to a large commercial transport). It may be reasonable to assume it's valid, then go collect your own data, with an test article instrumented in an "nz" accelerometer, a "loads monitoring" system.
But my experience is with metallic structures, and you've mentioned your wing is composite. I know composite primary structures is different (and no doubt others will post on it). I understand the most important thing is "BVID" ... possibly Barely Visually Inspectable Damage. I think you need to do a tap test or some other inspection of the test specimen (and use your FEA to determine the most likely places for a disbond) to determine if damage has occurred. I would rather continue the test until you get a failure, but you can always assume failure has occurred and you're back to the two previous points. But a single load test may be less valid for composites, I've heard that they tend to be more sensitive to low loads (high cycle frequency).
Another thing about composites is vulnerability to UV light, generally environmental effects
"Hoffen wir mal, dass alles gut geht !"
General Paulus, Nov 1942, outside Stalingrad after the launch of Operation Uranus.
SWComposites
Aerospace
Oh, don’t bring the BVID madness into the unmanned UAV world. Just don’t. No.
To convert test cycles to flight hours, you need to establish a baseline flight mission (or multiple ones if very different), then determine the number of cycles per flight and the number of hours per flight. Highly depends on the flight mission. Just a 1g loitering around mission, or lots of high g maneuvers??
Composites (well, carbon fiber composites loaded in-plane) are sensitive to high fatigue loads, not low loads at high cycles.
And, properly painted composites are not sensitive to UV. Just don’t expose the bare epoxy resin to UV.
To convert test cycles to flight hours, you need to establish a baseline flight mission (or multiple ones if very different), then determine the number of cycles per flight and the number of hours per flight. Highly depends on the flight mission. Just a 1g loitering around mission, or lots of high g maneuvers??
Composites (well, carbon fiber composites loaded in-plane) are sensitive to high fatigue loads, not low loads at high cycles.
And, properly painted composites are not sensitive to UV. Just don’t expose the bare epoxy resin to UV.
well SW has "composite" in his handle, so he knows more !
so would loads monitoring be the "best" approach to determine flights equivalent to cycles ? I'm thinking this is very low altitude so ESDU 69023 may not be accurate.
On one level ... why bother with any sort of accuracy ? Is this just a marketing figure, then count one cycle = one flight (particularly if limit load) ? Is this a certification requirement ?? (then see what your authority wants) You're in Eastern Europe, yes?
Have to say 20000 cycles isn't much ... 100000 is more like it for me. "BVID" may be "madness" but how else to assess the propagation of fatigue failure ? Your FEA is showing you the most critical parts of the wing, inspect there.
Maybe (??!!) I'm using "BVID" incorrectly. I don't mean in the certification sense, or in the fleet maintenance sense, but rather in inspecting the test specimen.
"Hoffen wir mal, dass alles gut geht !"
General Paulus, Nov 1942, outside Stalingrad after the launch of Operation Uranus.
so would loads monitoring be the "best" approach to determine flights equivalent to cycles ? I'm thinking this is very low altitude so ESDU 69023 may not be accurate.
On one level ... why bother with any sort of accuracy ? Is this just a marketing figure, then count one cycle = one flight (particularly if limit load) ? Is this a certification requirement ?? (then see what your authority wants) You're in Eastern Europe, yes?
Have to say 20000 cycles isn't much ... 100000 is more like it for me. "BVID" may be "madness" but how else to assess the propagation of fatigue failure ? Your FEA is showing you the most critical parts of the wing, inspect there.
Maybe (??!!) I'm using "BVID" incorrectly. I don't mean in the certification sense, or in the fleet maintenance sense, but rather in inspecting the test specimen.
"Hoffen wir mal, dass alles gut geht !"
General Paulus, Nov 1942, outside Stalingrad after the launch of Operation Uranus.
- Thread starter
- #5
Thank you both for your comments. We just did a trial test between 2g and -1g. The unmanned aerial vehicle can reach a maximum of 2.5g in the Z direction. We detected this in the simulation. Not 20000 cycles, we found 200000 cycles. Maybe there are points that are invisible to the naked eye. We are trying to do it with our own means.
I just want to know this. What kind of result can I give in terms of time based on this cycle?
Must I look at the ESDU?
I just want to know this. What kind of result can I give in terms of time based on this cycle?
Must I look at the ESDU?
"Must I look at the ESDU?" ... no, of course not; you don't Have to do anything. what we're saying is you need some logic to relate cycles to flights. That "logic" can be a statement ... "assume 1 cycle (+2g to -1g) is equivalent to one flight". Only problem is if someone challenges you on it ... "what data do you have for that assumption ?"
Ok, 200,000 cycles ... that's much better. That is probably conservatively equivalent to a safe life of 20,000 flights. I notice that I pitched you "safe life factor" but you didn't (apparently) pick up on it. This is an important parameter with fatigue analysis.
You could possibly say a safe life of 100,000 flights, but clearly much less "safe". again, if a marketing number ... go high (depending upon what it could cost you). If a certification number, then you'll probably need more than an internet discussion !
"Hoffen wir mal, dass alles gut geht !"
General Paulus, Nov 1942, outside Stalingrad after the launch of Operation Uranus.
Ok, 200,000 cycles ... that's much better. That is probably conservatively equivalent to a safe life of 20,000 flights. I notice that I pitched you "safe life factor" but you didn't (apparently) pick up on it. This is an important parameter with fatigue analysis.
You could possibly say a safe life of 100,000 flights, but clearly much less "safe". again, if a marketing number ... go high (depending upon what it could cost you). If a certification number, then you'll probably need more than an internet discussion !
"Hoffen wir mal, dass alles gut geht !"
General Paulus, Nov 1942, outside Stalingrad after the launch of Operation Uranus.
SWComposites
Aerospace
Im going to say it again, you need to define a) they load cycles per flight, b) the length of each flight.
One extreme might be to say that the max load cycle occurs every 1 minute of each flight. So divide 200,000 by 60 to get hours. But the max load cycle might happen more or less frequent than once per minute.
And if this was manned, the 200,000 test cycles would first be divided by a scatter factor of 5 (or higher). For a UAV, all depends on economics and risk.
One extreme might be to say that the max load cycle occurs every 1 minute of each flight. So divide 200,000 by 60 to get hours. But the max load cycle might happen more or less frequent than once per minute.
And if this was manned, the 200,000 test cycles would first be divided by a scatter factor of 5 (or higher). For a UAV, all depends on economics and risk.
- Thread starter
- #8
SWComposites
Aerospace
No, that is the time period of the test cycle.
What you need, or must estimate, is the actual usage flight profile of the UAV.
What you need, or must estimate, is the actual usage flight profile of the UAV.
You need to relate the test (200,000 cycles) to the UAV's operation. You want a relationship like "1 test cycle (+2g to -1g is a pretty severe cycle) is equivalent to 1 hrs flight time (or 1 flight, with an average time of 1 hr)". You get this either by simply saying it or by doing analysis of how you expect the UAV to operate (this is the ESDU I talked about, but thee other approaches, that determines the expected gust and manoeuvres that the UAV will experience, but like I said I don't think the ESDU is applicable to your operation) or you do "loads monitoring" (put an accelerometer on the UAV and go fly representative flights). From this you'll end up with a fatigue load spectrum of load 1g +- deltag at different frequencies, and you then do a fatigue analysis to determine the number of test cycles equivalent to a flight (is it 1, 10, 0.1). And you then apply a "safe life factor" to the result to get a safe operating life for the UAV.
This is a lot of work, maybe it is enough (as a UAV ... how big ?) just to state a life of 20,000 flights. How do you expect the UAV to operate ? 1/day ? so 20,000 flights is 50 years ... seems plenty ?
"Hoffen wir mal, dass alles gut geht !"
General Paulus, Nov 1942, outside Stalingrad after the launch of Operation Uranus.
This is a lot of work, maybe it is enough (as a UAV ... how big ?) just to state a life of 20,000 flights. How do you expect the UAV to operate ? 1/day ? so 20,000 flights is 50 years ... seems plenty ?
"Hoffen wir mal, dass alles gut geht !"
General Paulus, Nov 1942, outside Stalingrad after the launch of Operation Uranus.
- Thread starter
- #11
- Status