Problem with Modeling Truss in Staad Pro. 5

[WarrenG701]

Hello,

I am having an issue with a truss staad model. It is a simple 2d model of a single span monopitch timber truss supported at either end by bearing walls. The span is 38'-8" and depth is 12'-0". Top and bottom chords are (2) 2x6's, and vertical and diagonal web members are 2x4's. The bottom and top chords are sandwhiching the 2x4 web members. So there is one 2x6 chord on either side of the web members for all web members at both the top and bottom chords. I have provided pictures attached.

My loading is as following. I have applied a nodal point load at the second vertical web member of the truss due to the to the truss supporting a higher roof at this node. Additionally, I have applied a partial uniformly distributed load for the remainder of the truss due to the truss supporting a lower roof. (See the attached dwgs for the cross section of the truss and building/roofs). All of the web members have their ends pinned, so I have released all of the Mx, My, & Mz for all of the web members. The problem is that there is no resulting axial force in any of the web members or top and bottom chords, which of course is not possible. However, the overall deflected shape appears to be as expected.

The two things I am looking for help with are the following:

1)I am not sure if I release the ends of the top and bottom chords because I want to model them as continuous.
2) Why am I not getting any axial forces in any of the web members?

I cannot work out what's wrong. I have also attached the model for anyone who is would like to review it.
I am relatively new to staad so I'm sure it is something simple that I am overlooking. If anyone can offer some advice it would be greatly helpful.

Thanks
-Warren

Photo of Truss Construction:

https://files.engineering.com/getfile.aspx?folder=7b925828-e8b0-47eb-8e25-6ae9f26f88dd&file=024.JPG

Photo of High Roof/Low Roof:

https://files.engineering.com/getfile.aspx?folder=4da85f9d-b8a4-4d31-a76b-b86996b5e232&file=014.JPG

Exist. Truss Dwg 1:

https://files.engineering.com/getfi...2-4218-bd01-451e19d277e5&file=Truss_Dwg_1.jpg

Exist. Truss Dwg 2:

https://files.engineering.com/getfi...5-446b-a3d5-3eb177cfebff&file=Truss_Dwg_2.jpg

STAAD Model:

https://files.engineering.com/getfi...19e-ad9b-bc7dfeaf8227&file=Existing_Truss.STD

 

[Settingsun]

I'm not sure that I see the same analysis results as you see. The .std file you attached was missing section properties for the top & btm chords when I opened it and wouldn't run until I applied those properties.

After doing that, the model is unstable in several ways. These would be problems in any software, not just Staad. The analysis ended with 20 warnings and 7 notes. Always read those when the software raises them. Some other software would have raised these as errors rather than just warnings.

Problem #1 is global instability. You have a truss with only two pin supports at the bottom. Nothing stops it from falling over, just like a bicycle.

Problem #2 is local instability. You have released Mx, My & Mz for all web members. Nothing stops them spinning around their own axis (the Mx release at each end). Even though you haven't applied any torque to those members, it's still detected as instability.

 

[WarrenG701]

Stevehh49

Firstly thank you for taking the time to review my model. I am going to review my chord properties more closely. Also I will read the warnings and notes again. As far as the supports do you have a suggestion as how the supports should be modeled ? I know usually a truss is pin and roller connection. I am not sure on how to model this on Staad exactly.

 

[jayrod12]

I don't know staad, however you would put just a vertical restraint at one end, and a vertical and horizontal restraint at the other.

Also, in a note to steveh49's warning about the global instability. At the top of the truss you would put a horizontal reaction into the page to fix the global stability issues. That of course assumes there's something there in real life to stop the trusses from rolling over.

 

[BAretired]

Using pin supports each end of the truss is telling the software that the bottom chord does not change length, which is obviously wrong. Instead, use pin/roller supports (I don't know Staad, but there must be a simple way to do that).

The end diagonal in Photo #1 is a compression diagonal with a length of nearly 12', but is unbraced. If it is a 2x4, it is too slender. Other web members should be checked as well.

BA

 

[dozer]

Everyone has missed a fundamental problem here. Except for member 2 your web members are not connected to the chords. They are just floating in space.

To fix, select all members then hit the "Intersecting Beams" button in the Geometry ribbon. Select "Intersect Selected Beams" and accept the default tolerance.

There were some other good points made above but until you fix this you're just spinning your wheels.

 

[BAretired]

That's interesting! We should not be expected to find an error of that kind, as we are not familiar with Staad. This illustrates the danger of people using software which they don't understand. Something which I have harped about from time to time in the past.

Another thing which could be a problem...the value of E in the excerpt below is given as 230400 in two places and 230.4 in one place. Inconsistent units will tend to throw the results out the window too.

This approach to structural engineering is wholly unsatisfactory for reasons which should be obvious to all.

BA

 

[BAretired]

It would make much more sense for the OP to calculate the truss by hand methods, assuming pin joints throughout, with pin/roller supports. Forget about the continuity of the chords. They won't make that much difference. Focus instead on the significance of the results obtained by approximate methods.

BA

 

[WarrenG701]

@Dozer

Thank you very much, this solved my issue.
I was able to get the correct outputs after that.

 

[WarrenG701]

@BAretired

I would normally calculate these kinds of forces by hand. The problem is the client for the project is looking for a solution and a quick turnaround to issue repair sketches to the failing truss. Calculating the forces by hand doing the joint method would just take too long and is rather tedious. With all due respect, I have a full understanding of what’s going on structurally. It’s more so modeling in STaad that I require more practice with.
However, thanks for your input.

 

[BAretired]

@WarrenG701

I agree that you need more practice in Staad modeling. I can't help you with that as I do not know Staad, but, given your lack of understanding of the program, I am surprised that you would use it for a "quick" solution, to repair a failing truss. I do not consider hand calculation of a simple truss to be "tedious" at all, probably faster than inputting the data into Staad.

You have carefully avoided addressing the other points which have been raised.

BA

 

[Settingsun]

You generally aren't looking to save time the first time you try a new way of working. It's the subsequent times that see the benefit. I find Staad slower to use than several other programs but still quicker overall than hand calcs, much so as complexity increases. Once you've got the model set up, you get deflection at no extra effort, you can run the what-if (or client change) cases for very little effort, you don't have to file the calcs (if you did them on paper). It adds up.

 

[dozer]

@WarrenG701

When you say my recommendation solved your problem I hope you don't mean that's all you did. There are still other issues with your model, it's just that this was just such a basic problem it had to be fixed. You've got the Young's modulus of the web members 1000 times those of the chord members. It looks like the web members are the correct ones. BTW, you don't have to define separate material for members with different cross sections. You can have one material definition for them all if they are the same material. The program was trying to warn you something was amiss. That's why you have to pay attention to the warnings. Also, if you would have looked at the midspan deflections of the chords you would have realized something was horribly wrong.

Also, you have instabilities all over the place. You said this is a 2D problem yet you defined it as a 3D one. That's what the keyword "SPACE" does on the first line of the file. You could have used the keyword "PLANE" then you wouldn't have to worry about stability in the Z direction. In actuality, a lot of times you can get away with being sloppy and the problem with still solve accurately but why risk it? You should get rid of all your stability problems before you trust the results.

As far as the structure itself, I'm getting that longest diagonal (member 4) has a compressive force of 3.5 kips for D+S. I don't know anything about designing timber but I can use Euler's buckling formula. I get the critical load (not the safe load) is only 0.78 kips. How is this thing even standing? Yeah, I know it probably hasn't seen it's design load plus it found another load path.

I'm sorry to pile on but BAretired makes some good points. In the time you spent going back and forth in this forum you could have designed this truss a dozen times. STAAD or most any structural analysis program is a good tool but I don't think you want to learn by yourself on such a critical structure. Do a hand check and be absolutely sure you understand how this is behaving.

 

[Settingsun]

A structure you can analyse by hand is perfect to learn on IMO. Questions were raised because the results weren't as expected, exactly as should happen. Not every job will be as simple analysis-wise; he'll need to know how to use the software eventually and now some of the traps have been sprung when they're easily visible.

 

[MSUK90]

WarrenG701, you may have already sorted out your design and all the suggestions were great. However, I have few questions out of your actual query:
1) How did you modify those scanned papers digitally(any specific software)?
2) When did the truss actually built? All those seems hand drawn sketches and looks great to see.

 

[Liam Nesson]

Bentley created 2 excellent books for Staad pro, probably because their help tab is not very user friendly. Definitely worth buying.

Having an experienced engineer show you how they trouble shoot errors is a great way to figure out future problems like this for yourself. The deflected shaped was off the page, discontinuity or poorly placed releases are often the culpret.

Use the support "fixed but" and release whatever you dont want.

Housekeeping might have already been mentioned. For a small model like this maybe no big deal, but on larger models its important. Dont release each memeber individually, as this can create huge amounts of checking, create a release and apply it to groups of members. There is a truss memeber designation that is simpler to apply in this case.

Personally very quick hand checks should happen, even to show the model is in the ball park.