RISA 3d vs Hand Calcs Deflections 2

[MKarr]

Does RISA take something into consideration when calculating deflections that simple hand calcs do not? I've had other projects where hand checks have yielded similar results but I always chalked up the difference to 3D effects. But for my current project an issue caused me to run a beam simply supported by itself and the results were pretty far off from my hand calcs.

I ran a simply supported W14x26 spanning 20' with an evenly distributed live load of 1.25 klf. Hand calcs yielded a deflection of 0.633" and RISA is generating a deflection of 0.792". Just a simple live load deflection.

Thank you in advanced.

 

[Celt83]

likely shear deformations

 

[sandman21]

You have the beam set to lateral under design and not gravity. Which will reduce the stiffness based on the DAM analysis procedures.

 

[TehMightyEngineer]

RISA includes shear deformations but more likely is you're getting a flexural adjustment to stiffness (per Direct Analysis Method).

Per JoshPlum from a previous eng-tips post on this very subject:

Go to the Codes tab of the (Global) Model Settings. If you're using AISC 13th or 14th edition, then there is a flexural adjustment to stiffness (per Direct Analysis Method). This is approximately 20% and is intended to decrease the flexural buckling strength and column frames. In order to better approximate the inelastic buckling that steel frames actually experience. Though this is not intended to be used for service level deflection calculations.

Professional Engineer (ME, NH, MA) Structural Engineer (IL)
American Concrete Industries

https://www.facebook.com/AmericanConcrete/

 

[MKarr]

Thank you all. Turning the adjust stiffness off on the individual beam gave me my hand calc answer, and turning it off globally gave me the correct answer minus some 3D effects. So basically you would want to leave that on for strength calcs and then turn it off when checking service deflections?

 

[dozer]

MKarr, I ran it with a competing program and got a deflection of 0.652". I then flipped the switch to ignore shear deformation and got 0.633", your exact theoretical value. You keep talking about "3D effects" which I'm not sure exactly what you mean but I think it is the shear deformation that accounts for the slight difference. I've never used RISA other than playing around with a demo version but I would think it would have the ability to turn off shear deformation. You want to do this just to compare to textbook answers. There is no reason to do it in an actual design.

 

[MKarr]

By 3D effects I mean other members deforming causing additional deformation of the member in question. Some of which is accounted for in the supporting members and is easy to take out. Other engineers in my office use the same term and I assumed that's also what they meant but not entirely sure.

 

[271828]

0.633 / 0.8 = 0.791

It's the DM adjustment that some mentioned. Probably a very slight discrepancy from shear deformations, also.

 

[271828]

"By 3D effects I mean other members deforming causing additional deformation of the member in question. Some of which is accounted for in the supporting members and is easy to take out. Other engineers in my office use the same term and I assumed that's also what they meant but not entirely sure."

That's a little scary, to be honest. It sounds like an office-wide "mystery factor." We need to understand behavior precisely.

 

[MKarr]

I agree. And unfortunately I'm probably the heaviest user of RISA in my office and never really got any formal training on it so I'm kinda learning as I go.

 

[TehMightyEngineer]

so I'm kinda learning as I go.

At least you're learning and not just letting the black box give you an answer without verification. Sounds like you'll become the office RISA guru in short order.

Professional Engineer (ME, NH, MA) Structural Engineer (IL)
American Concrete Industries

https://www.facebook.com/AmericanConcrete/

 

[MKarr]

I try.

 

[BAretired]

Checking deflection by hand, I get 0.633" for bending and 0.021" for shear. This would result in a total deflection of 0.654" considering bending and shear together. Shear deflection accounts for only 3.3% of bending deflection.

I have never used and am not familiar with the DAM provisions, but it seems to me that, whereas they may be justified in a frame, they are not justified in a simple beam calculation.

BA

 

[TehMightyEngineer]

You are correct BA. They're intended for columns and specifically members with significant P-delta effects. I'm assuming RISA defaults to including the stiffness reduction provisions from the DAM by default because it's conservative and to avoid people erroneously forgetting to apply them.

Professional Engineer (ME, NH, MA) Structural Engineer (IL)
American Concrete Industries

https://www.facebook.com/AmericanConcrete/

 

[BAretired]

Well, I should probably keep quiet since I don't know a DAM thing, but if the provisions are intended only for columns in a frame, why would RISA apply the stiffness reduction to a beam?

BA

 

[TehMightyEngineer]

RISA isn't clever enough to tell on its own that the member you're inputting into the program is a floor joist or the beam in a special steel moment frame. Heck it can't even tell if it's a beam or a column on it's own. So, it's simply a matter of telling RISA how to treat each member.

My understanding is it defaults to the most conservative option. That way, if someone is just treating it as a black box it is at least accommodating the most conservative option. Plus it's only a stiffness reduction so aside from deflection the beam should still be sized appropriately either way.

I seem to recall RISA is also slightly conservative in it's application of the stiffness reduction to make it easier for the software to comply with the direct analysis method.

BA, if you want to learn about some of the failings of the effective length method (and thus what the direct analysis corrects via computer modeling) I highly recommend this lecture: [URL unfurl="true"]https://www.aisc.org/education/continuingeducation/education-archives/so-you-want-to-use-k-factors-do-you-the-effective-length-method-vs.-the-direct-analysis-method/#.Wd-6IDCX1PY[/url]

It's DAM good.

If you want a lecture specifically about DAM and how it's applied in the code this is good (but dry):

https://www.youtube.com/watch?v=iOH6kaXX3Mk&feature=youtu.be

Edit: I should also have noted that it's not just for columns but also for anything that's supported by columns. Thus, beams in moments frames will also see their load increased by the P-delta effects in the DAM. Primarily it's going to be columns that have the most increase though.

Professional Engineer (ME, NH, MA) Structural Engineer (IL)
American Concrete Industries

https://www.facebook.com/AmericanConcrete/

 

[DETstru]

DAM....

RISA doesn't have any automated way to distinguish between columns and beams other than specifically being told to by the user.

The DAM provisions apply only to members that contribute to the stability of a building. That can be a beam if that beam is part of the lateral load resisting system.

If you designate a steel member as "Lateral" (Hot Rolled Steel Design Parameters), and turn on the "Adjust Stiffness" setting (Global Model Settings), the requirements of C2.3(1) will be applied to that member.

If you also designate that steel member as a "Beam" or a "Column" (Section Sets), the requirements of C2.3(2) will also be applied to it.

 

[DETstru]

I see TMH already chimed in but just so I'm not being pointlessly redundant I'll add to the conversation about the DAM:

DAM requires P-Big Delta and P-Little Delta to be accounted for.
RISA can handle P-Big Delta simply by turning it on in the Load Combinations table.

P-Little Delta is not inherently calculated within RISA. RISA's P-Delta calculations are based solely on nodal displacements. So if you have columns and beams with only nodes at their ends, RISA only is aware of displacements at member ends when it determines P-Delta. This is all that is needed for P-Big Delta anyway.

To determine P-Little Delta, it needs additional nodes placed along the length of the elements to calculate the P-Little Delta effects. You can easily test this:

Model a single steel cantilevered column with a fixed base and free top. Preferably something not so stiff.
Add a large axial load and a small lateral load at the top of the column.
Run the analysis without any P-Delta.
Your moment reaction at the base will be exactly the lateral load multiplied by the height.
Turn on P-Delta in the Load Combinations and run the analysis.
Your moment will be higher due to the P-Big Delta effects.
Now add a few additional nodes along the column height and run the analysis.
The moment reaction again goes up even higher because of P-Little Delta effects.
Add more nodes... higher moments... etc.

RISA has published validation of the P-Little Delta calculation somewhere. I believe it also states how many nodes are required.
I'll dig it up and post it if I can find it online.

 

[Shotzie]

Verification Problem 12 speaks to this a bit as well:

https://risa.com/documents/risa-3d/RISA-3D_Verification_Problems.zip