Plate Bending FEA

[HardyParty]

I have a problem similar to the question in thread507-232128 and am looking for some advice.

I am trying to design an embed plate for 98k load and am having trouble figure out how to size the plate for bending. The 98k load comes from a beam which is attached by an all welded double angle connection (and has an eccentricity of 3"). I understand that this eccentricity causes bending in the plate. The plate has 9 studs and I am trying to figure out how it will bend in between the studs.

I made a finite element model and modeled each stud as a pinned connect. The contours match the stress I would expect, I just do not understand how to use the output. The attachment has a drawing as well as the plate forces I derived from the FEA. How would I use these to determine the thickness of the plate I need?

Thanks.

 

[Structures33]

I don't know what program you are using but typically, Mxy is the piece that will account for torsion and should be added to the Mx and My. In order to do this, you look at the contours and locate the points of max additive moment for Mx + Mxy and My + Mxy (you will have + and - for each for a total of 4). I use RISA and their tech support has told me to use the contour values (not those in the plate forces table).

Next, you will take the max absolute value of Mx and My and design the plate thickness to be able to resist that moment. You use the max absolute value of Mx and My because you are going to assign a consistent thickness for the plate. If this were concrete you would need to analyze separately to determine the r/f in each direction, top and bottom. I suppose to design the plate to resist the moment you need to assume a short "beam" section between the studs.

You also need to look at the reactions at each of the studs to design the stud sizes and embeddment.

 

[HardyParty]

Structures33:

I am using Risa as well. By looking at the contour stresses, do you mean the values in the small legend at the top right of the screen? This legend is called "Plate moment x, kft-ft." The maximum in this chart is 11.1 (rather than the 10.4kft-ft from the table.) Why would these values be different? (Perhaps this is something I should ask the RISA tech support.)

Once I have the maximum values from bending and torsion, let's assume is 13kft-ft. What would I used to determine the capacity of the steel plate? Of course I would want to use the section modulus:

My=FySx
Sx=bd^2/6

But what would be my "b" value? Would it be a foot since the force is in kft-ft? Would it be dependent on my stud spacing?

 

[rb1957]

1) wouldn't you worry about the welds in bending ?

2) personally i worry about someone making statements like "I just do not understand how to use the output." and "do you mean the values in the small legend at the top right of the screen?". this doesn't portray someone who knows the software they're using. i'd suggest cracking the manual and trying some simple test cases.

Quando Omni Flunkus Moritati

 

[JoshPlumSE]

There are times when plate contours of Mx and Mxy be the values that you want to look at. Certainly they give you a good idea of the overall behavior. Though structure33's suggestion that we use the plate contours in lieu of the plate spreadsheet force results is not something that I would often recommend. The spreadsheet values are the best values for the natural center of the plate (an average of the values at the Gaussian integration points). If your design needs to be based on the moment at the EDGE of a plate, then the contours use an algorithm to project the forces from the Gauss points out towards the edge. But, this is inherently an estimate at these locations.

Note: I tend to use the Mx + Mxy value more for concrete than for steel. Even then, I often prefer to use the plate corner forces or the "internal force summation tool". For steel, my tendency is to look at the Von Mises stresses instead. This is especially true whenever my primary concern is with the yield limit state. Von Mises stresses represent the "maximum energy of distortion" and are the easiest thing to compare against yield/Safety factor.

 

[HardyParty]

JoshPlum:

I read on the Risa reference guide that a Von Mises stress of above 36ksi would represent yielding of a 36ksi plate. Does this translate to mean as long as my Von Mises stresses are under 36ksi, my plate will not yield?

I understand this is very complex and gets into FEA theory and methods. I am trying to understand it to the best of my ability. Please forgive my ignorance. This is just such a big force, its not something I want to overlook.

People design embedded plates all the time, I'm not sure how most people go about this!

 

[delagina]

cant you compare stresses instead?
maximum stress you got from risa using whatever thickness you assume then compare to 0.75 Fy which is the stress capacity of plate using ASD 9th edition.

 

[Structures33]

I think it's probably a good idea to talk it through with RISA tech support. They understand their product and how it should best be used.

JoshPlum - to clarify - using the plate contours in lieu of the table forces was not my suggestion. As I stated, RISA tech support has directed me to do that - and they have done so on more than one occassion. I agree you need to consider where you want your plate forces.

However, that leads me to another point about modeling plates. There are some pretty particular rules for plate modeling that need to be followed. For example, the thickness should not be more than a 1:2 ratio to the dimensions of the plates - RISA recommends 1:1. When you model with a 1:1 ratio for something thin like a steel plate, your contours should be pretty accurate. You also need to be careful about the local axis of the plates - they all need to be aligned in the same direction in order to obtain accurate results from the contours. Also, the boundary conditions of the plates need to be consistent with what you are modeling. RISA suggests releasing the moment in the plate corners where they will be attached to a support if you want to engage that support. Otherwise, generally speaking, they will be fixed at the corners.

Von Mises can be confusing to me in RISA and I recommend having tech support assist.

Regarding your question about the section modulus - I think the important aspect of the analysis is to be consistent with the values in terms of the "b" and the units for the moment.

 

[JoshPlumSE]

Structure33 -

Actually the Max thickness to min width ratio that we recommend for RISA is 3:1. That shouldn't come into play for this model as you usually only get that thick for concrete slabs. But, I don't want the other number floating around as it is a bit more restrictive than necessary.

HardyParty -

The Von Mises stress is probably the best stress to compare to yield. At least as far as the RISA output is concerned. It's directionless, so you don't know if its a compression, tension or shear stress.... It only tells you how your plate compares to yield. So, it can't be used in every situation. It certainly can't be used when you are concerned about compression buckling of a tank or that sort of thing.

I suppose I could also point out that Von Mises can be overly conservative when compared to plastic design criteria (where you have to yield the whole cross section rather than just the extreme fibers). But, at that point, we're really splitting hairs.

 

[rowingengineer]

Just need to correct one part of this thread mxy dose not account for torsion, mx or my would have the torsion component included.

http://www.nceng.com.au/

"Programming today is a race between software engineers striving to build bigger and better idiot-proof programs, and the Universe trying to produce bigger and better idiots. So far, the Universe is winning."

 

[WillisV]

Use a 3/4" pl and move on.