Discretization of top and bottom chords for truss

[sonyablade2010]

Hi,

I'd like to know which is the best approach to model the top and bottom chord for trusses as unsplitted sole member or as a splitted member (from intersection points of diagonals and vertical members ) ?

Designed as splited member, normally I get all section and stability check passed, but if I design it as unsplitted member I get the instability due to deflection.

Regards,

[teixeiranh]

Can you please attach a pic of that instability?

The best way to do that is using a unsplitted element, at least for the general cases.

[StefanoPasquini6790 · Accepted solution]

Dear Spnyablade2010,

Surely the best way is to use the unsplitted bar, Robot is one of the few software in the world that give you this option and also the chance to create a superbar with a list of simple bars.....and that's great.

Everything depend by the design rule, specialy the buckling parameters.

Normaly the upper chord is compressed and have buckling restraint in z direction from the vertical and diagonal members of the truss. In y direction there are the purlins and sometimes the bracings.

The lower chord is in tension and have only a z direction buckling restraint.

This are only "general" rule, everything depend by your model. Please send it us

[sonyablade2010]

I completely agree with you that it would be a better to model it as unsplitted but couldn't it be designed as single pieces ? To meet the criteria of instability check probably I'd need well overdesigned chord. The chord in question is highlighted bottom chord as it can be seen from the attached pictures I also added the section check report.

Please find the files at the attachement.

I can't even believe how we design it at the past, not for the RSA but for other softwares e.g. early versions of SAP2000, there wasn't any hint on top/bottom chord are disconnected unless you specify it explicitly by splitting and connect them,  at least now they connect bars by some tolerance if they are close enough. 

Regards,

[teixeiranh]

The world is changing. And we appreciate it. I am a young engineer and I really cannot imagine myself working with the old fashion symbolic representation of, eg, rebar in reinforced concrete.

I guess that you need to adopt a larger steel section for that element. Or review the loads. It may seem odd that the element is compressed, but it is possible when you have upward wind.

[StefanoPasquini6790]

Wow....be careful with buckling....this cross section is absolutely wrong!!!

First of all, the dimensioning rule are not correct. You have to set the auto buckling restraint detection.

Second, I suggest you to insert some intermediate cross bracings between the trusses, for example in the column allignement. As I can see you have disaligned column axis and you can use them to brace you upper and lower chord.

I repeat my previous request: send us your model.

Greetings

[sonyablade2010]

Hi, 

Bascially I can cross the span with 100*50*5 box profile for top/bottom chord (S355JR) with splitted profile, on the other scenario if I have to model the bottom/top chord as single pieces then even 150*100*6.3 (S355JR)  is not enough, demand / capacity ratio is 3.61, there is huge difference.

For being more verbose: My loads are as follow and use full manual combination of those loads, except the fact that wind and snow never acting together in combinations

self load = calculated by RSA 

dead load = 40kgf/m2 

wind load = 64kgf/m2 

snow load = 75kg/m2 + snow dirft 

quake loads = computed by RSA 

Here is the project I played with cross sections but loading and combinations scheme is intact, so you may choose what ever profile you want, your help will be appreciated!!

https://www.dropbox.com/s/axdyjfi2q4uxzwu/Test_Project.rtd?dl=0

[sonyablade2010]

Hi,

After a struggle, If I solve the bottom and top chord as splitted bar elements I get 300*200*6 mm box profile which satisfies the slenderness and sectional check, but if I split the bottom and top chords 100*50*5mm box profile is enough. Surely I 'd prefer to use the unsplitted profile but there is drastical and dramatical section difference, It is more than 3 times bigger than 100*50*5mm, I'm looking for the more rational solution.

Regards,

[StefanoPasquini6790]

Sorry, the dimensioning of your bars don't depend by the fact that they are splitted or unsplitted!!!! The difference between the two solutions is due only to a "WRONG" definition of dimensioning rule.

If you can wait for tomorrow, when I will be in the office, I will give you the solution.

To better understand you issue, try to run a buckling analisis for an ULS combination, as for splitted model as for unsplitted model, and compare the multiplier.

If you build your structure you will surely see a buckling collapse, probably in assembly phase. I NEED TO STRONGLY SUGGEST YOU TO BE CAREFUL WITH THE BUCKLING PROBLEMS...Steel structures dimensioning are completely governed by the buckling.

Wait for me tomorrow....do not rush!!!!

[sonyablade2010]

Hi again, what do you mean by dimensioning rule is it span to depth (span/depth) ratio of truss?

[StefanoPasquini6790]

Good Morning.

I made a little investigation in your model and I found a lot of mistakes:

1) In this test model all the load combination are in wrong type, as you can see in the following picture they are all "FIRE" and not "ULS", I've edited manually the first rows...

2) The steel dimensioning rules are completely wrong...........I can't understand as you set it. If you watch my pic, I see that you don't set anything about buckling.....THIS IS ABSOLUTELY CRAZY!!!!

Attached to this mail the model with the correction in steel dimensionig rules and the definition of superbars, starting by a splitted bar model. Ypu can also find the dimensioning report for upper and lower chord as for wrong definition as for right definition, with the minimum profile that pass all the ULS check.

I think that you need a little bit more training, stay tuned in the forum and watch the videos in the webinar channel....Be carefull, this is not a videogame!

files: https://www.dropbox.com/sh/wzugrz9iqvwg9wx/AABDQfwnJ7HfXuQ0LZ_HBE-ua?dl=0

[sonyablade2010]

Hi Stefano,

Thanks for helping, I really don't know how but I've sent you the wrong and uncompleted project, sorry for that, in any case this projects are draft. For the quotation purposes I usually tend to solve the approximate and rough project at least to get the approximate sizes of profiles, then If I'm awarded for the project, I go with all the details, loading, combination etcc. type. 

Please refer to that project:

https://www.dropbox.com/s/49h8kgoj1f5btm4/Test_Project_3D.rar?dl=0

With that new project especially 3rd axis truss which has largest span 15.6mt (with splitted bottom chord ) is ok in every respect, but in the next 4th axis bottom chord is failing which has relatively smaller span 10.68mt (unsplitted) member. 

IMHO here is the clear difference between having the splitted and unsplitted members, so which one modelling option is correct? 

Regards,

[StefanoPasquini6790]

Hi,

here comes the model edited and updated, you have to check loads and load combination. Some notes:

1) Work first with your model without seismic loads and, when you have finished with the dimensioning add the seismic loads and combination and run the checks. Usually earthquake, specialy in only one story buildings, is not decisive!

2) I managed to adjust the geometry of the model and I created a list of dimensioning groups.....

3) Your model have seriously problems with buckling, I run a buckling analysis for a ULS case and have this multiplier:

0,52
0,62
0,70
0,83
0,84
0,92
0,92
0,96
1,05
1,14

If you think that the minimum multiplier must be equal to 10.....and you have 0.52????

4) You have to think about the cross section of the column, I would use HE section instead of 2 UPA,

My opinion, your truss height (0.7m) is too small to cover an almost 16 m span? The average ratio between span and height have to be higher then 1/15....

Insert purlins in your model, they are important for truss bucling restraining.

I do for you the best that I can...now is your turn!!

Cheers

[sonyablade2010]

Hi Stefano,

Thank you again,

• that span depth is really small so I'll try to increase it, at least I'll try to make tapered style starting from 80cm going to 50cm at the mid span, along with that fact I also found it usefult for other collegues as quick reference on geometrical paracticality and joist composition by Steel Joist Institute(SJI) based on the tentative data, it says that the max of 12mt span can be crossed by 508mm joist depth e.g. 20LH10.
• I'll try to brace the bottom chord horizontally rather than vertically, as you said top chord will be braced in any way by the purlins.
• Based on your post at the  Critical Buckling Length, as the other poster noted it is not totally mandatory if you perform 2nd order analysis it can be smaller, but question arises to which value? It 'd be a better to stick to your design suggestion (alpha_cr > 10) but I'm afraid that it will lead to increasing of overall rigidity, no matter what the truss geometries are.

Your help really appreciated,

Regards,