our 4 poster rig can reproduce a limited bandwidth. Could it be fine anyway for me?

[pietro82]

Hi all,

I have installed strain-gauges on the axle of a commercial vehicle to measure the vertical and the horizontal loading. I placed them in order to reduce the cross-sensitivity at lowest as possible. I need to reproduce the measured loads with a 4-poster rig. Unfortunately the maximum reproducible frequency by our rig is 20 Hz. I need to know if our rig can reproduce the loading history correctly, therefore I have in mind to compare the pseudo-damage of the loading history with the one of the filtered loading history with a low pass filter with a cut-off frequency of 20Hz. it is the right way to proceed, isn't? Otherwise may you provide me a better one?

Thanks

Cheers

Pietro

 

[GregLocock]

I think the problem you have is that (a) 20 Hz is a bit on the low side for a maximum rig frequency but (b) far more importantly 4 poster rigs are notoriously difficult to set up for rough road type events, as you really need longitudinal inputs as well.

However within those limitations yes your approach makes sense. Are you driving the hubs directly or do you have the vehicle sat on tires (in which case the 20 hz won't matter much as you are isolated by wheelhop)

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[pietro82]

Hi Greg,

Thanks for your reply. Our 4 poster rig drives the hubs directly. What is the best approach? I filtered few signals with low pass filters with different cut-off frequency and then I calculted the damage with Miner law: I expected that higher was the cut-off frequerncy and higher was the damage, but it wasn't always true. I read on this paper: Fatigue Life Modelling and Accelerated Tests for Components under Variable Amplitude Loads (you can download it here, that spectral filtering can lead erronuos fatigue life results due to filter roll-off and phase shift. So what is the best approach?

thanks

Cheers

Pietro

 

[GregLocock]

From my very much secondhand perspective, accelerated fatigue damage is best represented by a rainflow type analysis, rather than getting too het up about frequencies.

Driving the hubs directly gets rid of the wheelhop problem, but to get realistic fatigue lives we use multi dof excitation of each hub, at least 3 and maybe 5.



Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[pietro82]

Hi Greg,

The vehicle I have to test doesn't have any suspension system, just a pivoting front axle and a rigid rear axle. Yes the fatigue damage is better represented by a rainflow analysis, but I can't assess if I can get rid the frequency higher than 20Hz with this method.
I will excite each hub with a vertical excitation. In the paper before-mentioned it is stated:
"Successful fatigue editing is also possible by combination methods. As an example, frequency correlated damage based filtering can remove cycles whose frequency cannot be achieved and the resulting filtered multi-channel time series set can be further processed in time correlated damage editing to produce reduced time series as the first set of targets for iterative transfer feedback control."

Do you know any reference material of the "frequency correlated damage filtering"?

Thanks

Cheers

Pietro

 

[GregLocock]

Sorry, as I said, I don't actually have much to do with fatigue, so I am not up on the literature.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[pietro82]

Hi Greg,

thanks for your reply. Hopefully someone could give me a tip.

Cheers

Pietro

 

[pietro82]

Can't anyone help me?