Anyone designed joists in Risa? As in for fabrication? 2

[Hemi79]

Iv'e spoken to some local steel fabricators that claim they fabricate their own joists for many projects as opposed to going with an SJI certified manufacturer. I have an issue on a project where joists can't make it on time and locals are offering to fabricate. The joists are 65' lengths and 40LH10 were designated. This makes me very nervous and Im curious to starting a truss in Risa, eliminating the very first vertical and horizontal members on each side and trying to see what the joist members (chords/diagnols/verticals) should be. Anyone attempted this or know how common local joist fabrication is?

Thanks,

 

[JoshPlumSE]

I've seen people do it before with RISA. Though it has been some years since I really delved into one of their models. My thoughts:

1) You have to be particular about the unbraced lengths you're using. The SJI guys really push the limit.... like using a KL value less than 1.0 for web members. Maybe accounting for stiffening of the web at the connection to the chord. That's okay for them as they can back up their numbers with test data. I tend to be a little more conservative with my modeling.
2) If I recall correctly the SJI type joists tended to ignore some things like Torsional buckling of the chords. So, the RISA numbers ended up being a little conservative.
3) Bridging is always an issue (not with the model), but with construction. So, you should probably give a heads up to the construction crew that they should follow the stability recommendations from SJI even though it's not an SJI joist.
4) Do joist fabricators usually use 50 ksi angle members? I usually use A36 for angles.
5) For P-Delta / Direct Analysis you might have issues for the top chord buckling in the out of plane direction if you apply a notional load (or 'kink' it to directly model imperfections). But, this is because the transverse direction is likely restrained by deck or sub-purlins or something. Not sure how to demonstrate if the deck is stiff enough to truly restrain the top chord.

I'm sure there are lots of other considerations as well.

 

[BUGGAR]

Considering the requirements of SJI (good or bad), you will likely have to ignore SJI and stick with AISC and AWS (SJI doesn't necessarily follow AWS - look at some of their welds?!?).

 

[JoshPlumSE]

I agree with "Buggar", though I didn't explicitly state that in my original response. The others that I've seen do this sort of thing have done their code checking / design using the AISC code criteria.

My response was meant to clarify that by doing this, you can get relatively close to the values of the SJI tables. But, that you will likely be a little more conservative. Joist fabricators / SJI have years worth of experience and testing to backup some liberalizations on their code value assumptions. Also, they have a lot more quality control over the construction of the joist itself.

 

[Hemi79]

THanks for the info @BUGGAR and @JOSH! Running the design in risa I get crazy chord members. My top chord is an 8x8x.75" and that adds up to 5kip total weight of member. I obviously have some optimization to do as the weight of the joist by SJI certified manufacturers is approximately 20lb/ft. I currently have my verticals/diagonal at 2.5' O.C. I'll have to wait and see what they span their panels at and chords members they propose. I have modeled 2 joist with exact members and spacing, one with point loads over verticals and the other with the uniform loading through out the chord. The point load receiving model is far more critical than the other. I mean every member is an axially loaded member instead of a "beam". Again, THanks, I'll be posting as I get some progress.

 

[dik]

Can you not have the fabricator do and seal the drawings for the roof trusses based on stipulated loads and applicable code? That's what I would normally do.

Dik

 

[Hemi79]

@dik, in fact yes. That was my first response. However, the client is a longtime customer of my Boss. And they want to fast pace the project so we are trying to work with what we have. I got a full lecture about customer, sales, "wearing different hats" etc.. Seemed annoying at first but after it sunk in I kinda felt he was right and I could spend some more time trying to make things work. This is definitely out of my comfort zone engineering wise. THat's what sucks the most.

 

[dik]

Can you not explain to your client that there is an economy in having 'experts and specialists' provide the design for items they work with on a 24-7 basis? and, being experts, they are likely faster.

Dik

 

[BUGGAR]

Steel Joists are kind of like Metal Buildings (MBMA type). It's hard to make them "figure" and you can't duplicate them economically. Nor do you want to?

 

[dik]

Bugger: Concur... Had one once, where sprinklers were added... an extra 3 psf... some of the redesigned members actually increased... not unreasonable, but, unexpected.

Dik

 

[Hemi79]

So here is what I have so far in case you all can look at it. I added a 500 point load miscellaneous at center and a 550lb point load for mechanical equipment. They will have to use Diagonal Briding so I have accounted for that at 5 ft. on center along the span. Of course there will be 22 gauge deck over it as well. Looking at it now, I do feel its not as bad as I thought. Obviously pricey but doable.

Thanks,

 

[JoshPlumSE]

Your boundary conditions are pinned - pinned. Allow one of them to move in the X direction or your truss analysis will not be correct.... You'll know because you'll start to see tension in your top (i.e. compression) chord.

 

[Hemi79]

@joshplum, thanks for the help. Let me ask though, setting my x direction to no reaction, isn't that not realistic? Joists will take a weld at each end, transverse and longitudinal, yes?

 

[JoshPlumSE]

The way I usually explain it is this:

1) Run the model restrained in the X direction. What is the X direction reaction? In your case it is 38.7 kips for LC#3.
2) Run the model free in the X direction. What is the deflection at that spot? In your case it is 0.5" for that same load combination.
3) Considering the rigidity of what you're connecting to and considering that there will be some "slip" in the connection as well, if the support were subject to 38.7 kips, would it deflect 0.5"? Probably... If so, then it is easier to the boundary condition being free in the X direction rather than fully restrained.

Obviously, the reality is that truss support is somewhere between restrained and free. But, the traditional assumption is that it is free. And, when you really closely look at the numbers it usually behaves much closer to free than restrained.

 

[Hemi79]

@Joshplum, thanks for taking the time to explain this. I really appreciate it. Extremely helpful!