Poisson ratio appears to be changing based on application/relaxation of load? 2

[adammil1]

I was wondering if anyone could explain this behavior to me. The setup was I had a custom made fixture which was more or less a 3.75" high by 3" dia 6061 aluminum block put onto our compression testing machine here at work. I loaded it up to 14000lbs, and then unloaded it. Initially the ramp rate of the stand was rather low I want to say it was .05"/min going up and going down.

I had always thought that the poisons ratio would have governed the fact that given a particular force one would expect a particular deflection. Thus I would expect to see a single line.

However here’s a view of my initial run.

So thinking of it a little more the only thing that I could think of that may explain such behavior was perhaps the rate of load application would have some affect here. Perhaps it took some time for the molecules to recover or something. But as one can see on the plots of different rates of application that wasn’t the case.

So I was wondering does anyone know the cause for such behavior? I can’t be the first to have come across this.

I am pretty sure the machine itself is telling me the truth but you never know.

 

[adammil1]

Well it looked like one of the pictures didn't work out, so here it is again;

 

[CoryPad]

If the strain is parallel to the force application, then Poisson's ratio has no effect. Elastic modulus (Young's modulus) is the property that relates stress to strain. Compression testing, and the compliance correction for the fixturing, is always tricky. You need to look at alignment, angularity, etc. Plus, the accuracy of the force and strain measuring devices especially at the low values within their ranges.

 

[EdStainless]

How accurate is alignment of the load faces? And do they stay aligned under load?
I have done this with strain gages on four sides of the block under testing. If they don't match you have issues.

If you want this to be highly accurate you will need to first load to 20%, then unload to 5%, and then reload and measure. The play in the system will be a big factor if you actually relax the load.

= = = = = = = = = = = = = = = = = = = =
Plymouth Tube

 

[EdStainless]

Is the strain that you are measuring in the load direction (height) or is it lateral?
In the second case poisons ratio would come in to play.

Either way, until you have cycled the load a few times and gotten the exact same curves you don't have reliable data.

= = = = = = = = = = = = = = = = = = = =
Plymouth Tube

 

[adammil1]

Thanks for the replies. It looks like I confused youngs modulus with poisson's ratio this morning. None the less, I am indeed placing the cylinder in compression am trying to figure out where this hysteresis is actually coming from.

 

[mcguire]

The fact that you get an increasing modulus with load leads me to suspect there is a friction effect constraining lateral expansion.

Michael McGuire

http://stainlesssteelforengineers.blogspot.com/

 

[EdStainless]

When you cycle it three of four times, without going to zero in between, do the curves repeat?
Does the loop stay the same?
How are you measuring the displacement?

= = = = = = = = = = = = = = = = = = = =
Plymouth Tube

 

[Compositepro]

If you are getting any yielding in the specimen this is exactly what to expect. Are your results repeatable? 0.05"/min on this type of test is very fast.

 

[Ron]

Your load is only a little over 5 percent of yield, so I suspect measurement inaccuracies. In this range, all strains should be recoverable.

 

[Compositepro]

Are you measuring strain by cross-head travel? This will not be accurate. Friction and stretch in the drive and load frame will cause the non-linearity and hysteresis displayed in your curves. You must use an extensometer on the specimen to accurately measure modulus.

 

[adammil1]

The results so far have been very repeatable when tested from .001ipm to .1ipm as seen on the following graph that I had posted above;

http://files.engineering.com/getfil...e77-b313-566af32bf707&file=multiple_loads.PNG

The way the fixture is actually loaded gives me a contact pressure of 6900psi, so I am still quite a bit below yield.

Strain is being measured right now solely by cross head travel by the machine, I would have to do more research on the machine as to how it is actually measuring this but I would suspect they have a glass scale much like a Digital Readout on a machine tool? Friction was one of my first thoughts too as to what may cause this behavior however I would have thought the effects of friction would be more along the lines of slowing the movement of the material, and thus I would have expected a far greater dependence on crosshead speed to the curve shape than I see.

Ed, I haven't tried cycling the load yet with out actually taking it back to zero, what would cycling it back and forth without going to zero accomplish and what load should I take it down to?

 

[CoryPad]

Compositepro is correct, you need an extensometer for strain. The cross-head travel is not suitable for this purpose.

 

[Compositepro]

On a screw machine the rotation of the screw is measured as an indication of crosshead displacement. Any mechanical flexing, stretch, or torsional flexing of the screw will cause an error in the actual displacement. The load measurement is not affected so strength measurements will be correct.