Strain measurements - how to evaluate results?

[mechanicalDuck]

Hello,

I have a steel frame structure which is subjected to very large transient vertical and horizontal forces. I need to measure the maximum reaction forces in the bottom of the frame structure.

The loads on the steel structure are very large and occur due to discharging of large spring batteries.

I have applied strain gauges at the bottom of the frame's legs in order to measure the strain as a function of time.
The result is transient strain vs. time plots in which the strain oscilates greatly (and seemingly randomly) over the given time interval (< 1 sec).

Now, I first thought to use the maximum strain from these plots and convert it into stress and then into force in order to get the resultant reaction forces in the bottom of the structure. However, looking at the strain/time-plots I wonder if this is a good way to get reliable results.
Like I wrote, the strain varies greatly over this very short time interval and I wonder if just taking the maximum strain and converting it into force and stating that this is the maximum load at the bottom of the structure is correct?

Does anyone have any input on this problem? Should I somehow look at the total strain/time-curves and try to get an average reaction force at the bottom (and if so, what would be a good way to use the output data from the strain measurements to do this?) or should I go about the problem some other way?

 

[desertfox]

Hi mechanicalDuck

The way I see it I would use the maximum loads as this would give worse case situation.


desertfox

 

[KoachCSR]

I agree with desertfox. If you take the worst-case-scenario of your results, then that conclusion would be bounding for all other situations. For further conservatism, you may consider adding some margin (say 10%) to those loads following your calculations.

I agree with your stated method of calculating the force in the bottom of your structure.

Another consideration unrelated to the method of calculating the forces is sample size. I'd be sure that I took enough samples to accurately reflect the behavior of the frame under the worst-case loading condition.

 

[mechanicalDuck]

Thank you for your replies Desertfox and KoachCSR.

I have done exactly as you commented KoachCSR, i.e. I have enough test samples in order to get an accurate view of the loading.

Regarding taking the maximum loads to create the worst case situation: The maximum loads are just very very short transient peaks (a few milliseconds) and I am worried that by taking this value as maximum load would give me a very conservative load picture. Would a root-mean-square value of the plot be completely wrong or might there be some other way of determining a good overal load value?

 

[GregLocock]

Fatigue in response to a few cycles of large transient forces is a bit tricky. I agree that your approach is conservative. If your structure is steel then it will cope with quite astonishing multiples of its yield stress, briefly.

But there is no real reason to think that a factor of 0.707 is justified, tho I think it is a good place to start.

Incidentally welds are nothing like as robust.





Cheers

Greg Locock

SIGlease see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[KoachCSR]

mechanicalDuck, I take it you are referring to the effects of fatigue stress on your structure due to the cyclic loading. My previous understanding was that you simply wanted a snapshot, so to speak, of the maximum forces exerted.

If you are, indeed, looking to evaluate the effects of fatigue stress, then you may require the use of an FEA software such as ANSYS, particularly for the welded joints.

Was this structure originally designed to withstand this cyclic loading? If so, I would suspect that fatigue stress, provided there are no cracks in any welds or other such defects, would not be a major concern as the fatigue life of the structure should have been taken into consideration in the original design.