Effect of poisson ratio 1

[Rot_HVSC]

Due to going through a Dunning-Kruger drop in FEA confidence I'm starting to question some of the basics, like how important is getting the poisson ratio correct when using bricks and shells?

The Ausy steel code AS4100 states to use 0.25 but most literature I have read states poisson to be around 0.3. I know I could do a sensitivity analysis but with 3 + hour analysis run times, I can't justify the extra runs when our day to day hand calc's don't consider poisons ratio (treat everything a simple or encased beams)

Can I please have some advise on when changing the poisson ratio has significant effect's on anaylsis results. i.e. over 3 significant figures.

 

[FEA way]

Poisson's ratio is often treated rather lightly. But you could make some largely simplified tests if you are concerned about it. They should just approximately represent the type of geometry and loading that you usually study with FEA.

 

[3DDave]

Since it's a measure of the bulk compressability of the material due to a single-direction load the smaller value they use says the steel is more compressible - under tension it won't neck down as much and under compression it won't bulge out as much.

I don't see a steel that is that low, but ductile iron appears to be and cast iron is even lower.

As far as I can tell, increased Poisson's ratio means increase stress concentration factors, which makes sense. If an example part - a straight bar with side notches and a Poisson's ratio of zero - doesn't further pinch in when pulled then the stress won't change. Compared to a rubber band with a Poisson's ratio of .5, the edges of that notch are seeing a significant stress increase.

Using a lower than actual number might underestimate the concentration a little. The effect you'll see is going to be very geometry dependent - so the earlier suggestion to try simplified tests of the sort of geometry your are encountering is best. Also, check with the particular FEA supplier to see if they make any particular assumptions on the basis of the Poisson's ratio. They may not, but then I thought a major FEA supplier would handle Imperial units just fine and they weren't even close (yeah - a PCB resistor dissipating 1MW was seeing a rise in temp of around a milli-Kelvin. Worked as expected in SI units. The maker was surprised; I suppose they repaired the defect.)

 

[rb1957]

I'd be interested to ask "why 0.25 ? what research shows this ??"

Standards people don't just pick numbers, particularly a number contrary to common practice.

another day in paradise, or is paradise one day closer ?

 

[dik]

really good question; it's generally taken as 0.3 in any foreign codes I've been involved with. I should have added, it has a big impact on Shear Modulus.

Rather than think climate change and the corona virus as science, think of it as the wrath of God. Do you feel any better?

-Dik

 

[3DDave]

Poisson's is interrelated with a number of other factors, but those factors are often called out separately, hence my suggestion to ask the FEA maker what they use it for.

The most likely "didn't choose that number" reason is they copied it from something else and didn't pay close attention. Do formulas in that standard make use of it?

I found a reference book online (243 pages worth) on the subject of Ausy steel code AS4100 and it mentions the Poisson ratio as being .25 only in the symbols table and not any other time as far as I can tell. One would think if it was important to the document it would get more of a mention. Not being able to search for the Greek nu symbol may have skipped it, but Poisson wasn't mentioned anywhere else and the value for it didn't appear in any examples.

It's "Steel Structures Design Manual To AS 4100"
First Edition
Brian Kirk

It was on what looks like a legit site, but on the side of caution - you'll have to use google if you want it, mostly on worries about copyright.

 

[Rot_HVSC]

Thanks for the great answers every one.

3DDave at the office I do have access to a version that book so I'll be having a good read also of the standard. If any one wants I am happy to report anything interesting I find on this thread, just let me know. I guess it also best not to assume on how the FEA software uses inputs/or even how the software elements work (Shear locked 4 or 6 node plate element anyone?)

 

[NRP99]

Is this related with Poisson's original assumption of 0.25 for all solids? This is only reference I can relate to this value.

I am bit surprised by the "Standard" adopting Poissons ratio of 0.25 when almost all other structural codes are using 0.3 for steel which is inline with testing. Not able find any reason in the Standard itself. (Not looked hard enough though due to time limitation.) Nor find any explanation outside code.

Anyway, is that make any difference if I use 0.25 instead of 0.3? For 1% tensile axial strain of rectangular bar with axial tension, the lateral contraction strain would be 0.25% instead of 0.3%. This means the lateral strains and lateral stresses are less than 16.7%. For multiaxial loading, further less in multiples of 16.7%. So IMO, not conservative assumption.

 

[dik]

I suspect minor, but as noted, it has an impact on the Shear Modulus which may be of more significance with FEM.

Rather than think climate change and the corona virus as science, think of it as the wrath of God. Do you feel any better?

-Dik