Regarding modelling of Tension rod system of Skylight in STAAD

[Arbu]

Dear All,

I am trying to analyse and design a Tension rod system for horizontal skylight with tension rod at top and bottom. The concept is given by consultant. But I am getting difficulty in analysis and design. My question is which element I have to consider Either tension or cable while analyzing. If I am considering cable i cannot provide Pretension as it is tension rod only. If I am considering tension rod then deflection is infinity. I posted same question in STAAD forum but didn't got reply. Please reply if any one is having Idea. Drawing has been attached.

 

[Arbu]

Dear All,

I am attaching a STAAD file also. Please check.


Regards,
Sigma

 

[WARose]

I don't think you have any choice but to use a cable designation (and a Initial Tension) for the chord members. (The diagonals likely could stay as tension only.) Without that, this thing won't stand.

Even with that you might still get a stability warning for those vertical truss members (that have no releases and are not designated as tension members). But I would address the cable situation first.

 

[Shu Jiang PEng SE]

WARose is right. You have to specify chord as cable and give an initial pretension. The diagonal members have to be treated as truss members (can take both tension and compression)

 

[Arbu]

Dear WARose and Shu Jiang,

thanx for your reply.

some body ask me to provide camber in top cable and initial tension as 1 kN in all cables, but this also not working. Please check the attached model. I think without pretension not possible in this system. But when I am assigning a pretension then its showing stable (deflection < 100 mm) at a pretension greater than > 2000 kN with more than 250 warning of instability.

 

[klaus.]

I don't think this system will work like that ....
top chord and bottom chord are Tension (cable) members only ???
You would need very high pretensioning.....

 

[Arbu]

Dear Klaus,

yes top and bottom members are tension rods only.
I agree with you. Same results I am getting (very high pretension). But one consultant proposed this system and i have to verify. I will inform him that this will not work without pretension. I am new in such type of system that's why posted here.

As per my basic understanding about structural Engineering, to carry this type of loading there will be either flexural action or centenary action. As top and bottom members are tension rod flexural not possible. For centenary action it need to allow sag which is not advisable in skylight. The another option is High pretension which will lead very high lateral forces on column. I think better i will propose some other system.

Thanks once again for your quick reply.

 

[Tommy385]

Even if it will somehow work with cable designations you will have pretty big horizontal forces on you steel columns (they should work as cantilevers in that case...)
and if all of it still somehow works, you will have pretty large deformations, in combination with facade and roof glass elements maybe its not the best idea...

 

[klaus.]

yes...i agree ..this system will not work like that in reality
A tension member can also take compression as long as the compression force is lower that the pre tensioning force
in that case the tension member acts like a truss member
Also structures with cables only will usually have higher deformations.......

 

[ldeem]

If you are going to use the Cable command you need to carefully read the STAAD technical reference manual to understand how the command works. From my recollection the cable command changes the stiffness of the member but does not induce the tension force into connected members. You can get the tension force using the Pre/Post load command - read about this as well because the command is not intuitive.

I used to teach STAAD for REI and a little for Bentley and always advised students to very carefully read the technical reference manual when doing an advanced analysis and/or using commands they are not familiar with. It is very easy to misuse a command and get bad results - not bad calculations but Garbage In Garbage Out.

 

[WARose]

@ enggsigma, why even do this system at all? With some of these spans and loads......a joist/joist girder system would work pretty easily I'd say. (Vulcraft has a on-line catalog in case you are interested.)

 

[Arbu]

Dear Tommy385,

thanks for reply.
this fact I already discussed with our management that tension will be very high and column will not be able to carry. This problem is due to our estimation team proposed this concept and we got approval. When I started looking in detail design then I realized this system will not work.

Dear Ideem,

thanks for your reply and suggestion for cable member specification. Yes I agree there are lot limitation in cable analysis like we cannot use member load, lateral load, dynamic analysis, master and slave node etc.

Dear WARose,

client wants minimum obstruction for light from skylight. They are not allowing to use beam/joist system. But as the Cable system is not working I proposed another tension rod system with truss.

Regards
Enggsigma

 

[Arbu]

Dear All,

I proposed some what similar system but with combination of tension rod and trusses. In this one direction three trusses are running and other direction tension rod system is there. And the tension rods I had modeled as a cable having 5 kN pretension. There are around 180 warning of instability in STAAD regarding Instability. But I think we can neglect these instability Warnings as cable is loaded with lateral point loads(deflection in within permissible limit). Please check the attached model of STAAD. Please check and comment. First time I am using such a system thats why I need your reviews/comments. I checked one segment of cable between two truss manually I am getting almost same tension in cable as STAAD model. Please check.

 

[klaus.]

well i do not know Staad ... but please do not ignore any instability warnings if you are not sure where they come from
Working with tension and cable systems require some good knowledge....if you are new to this subject please be careful !!
Accidents might happen fast

 

[Arbu]

Dear Klaus,

Thanx for suggestion, I will check once again the warnings.

 

[Arbu]

Dear All,

one more strange thing I observed in this model. Its designing all the steel members for last combination only. I had put Perform Cable analysis and change after every load case then also this is designing for last case only. Can Any explain this why its happening.

 

[klaus.]

Again I do not know STAAD...but Usually you do not design the cables or tension rods by the static programm
Cables and tension rods are designed by the manufacture 'design' tables

You take the max tension force from STAAD and then choose the appropriate cable

 

[ldeem]

I tried running the STAAD model but it fails to solve. I tried with program version 20.07.11.90 (an older version) and 21.00.00.57. I haven't gotten around to installing the latest version. I would suggest changing to P-Delta analysis - that is easier to use and will help to show instabilities if the model does not converge. I would try both using the default convergence setting and multiple iteration setting - say 20 iterations. Treat the cables as tension rods and don't forget to add some load to simulate the tension.

To back up quite a bit why not use the truss idea you have but have trusses running in both directions. This would form a strong grid to support the load. Then between trusses you could design some lightweight frame to support the glass. Maybe it could be some sort of cable system like you are thinking of. The main STAAD analysis would just be the rigid trusses - this would avoid having the intermediate framing effect the design. Something like this would be much easier to analyze. Watch out for your out of plane unbraced chord lengths. Especially on the bottom chord where it could experance compression under strong wind uplift condtion.

You may need some kind of expansion joint in one direction as well.

I don't see any snow load so I assume you must be in a very warm area - very nice as I am expecting snow here in Ohio today.

 

[Arbu]

Dear Ideem,

actually the rods are tension rods only, for analysis purpose only I m considering it as cable with very less pretension of 5 kN. And I proposed initially both direction truss but Architect is not approving, he is saying too much trusses will be give a feel of industrial structure and this is a government office. As the structure is in middle east there is no snow.

One more strange thing I found in this model: I given design command for steel columns an truss only. But the STAAD is designing steel for last applied load combination No 1109 only while designing its suppressing all the remaining load cases. For tension rod I am taking maximum tension and comparing with capacity of tension rod. Is someone faced the same issue before?.

 

[WARose]

[blue](enggsigma)[/blue]

Dear WARose,

client wants minimum obstruction for light from skylight. They are not allowing to use beam/joist system. But as the Cable system is not working I proposed another tension rod system with truss.

With the truss girders you have in there now......why even do the tension members at all? Seems to me like you could leave what you have now and just have light joists making the span between the girders. The depth would not be changed (assuming of course your girder makes it as is).

[blue](enggsigma)[/blue]

There are around 180 warning of instability in STAAD regarding Instability. But I think we can neglect these instability Warnings as cable is loaded with lateral point loads(deflection in within permissible limit).

Never neglect instability warnings. They must all be resolved. (Quite a few will go away once you get rid of this system.)

[blue](enggsigma)[/blue]

Please check the attached model of STAAD. Please check and comment. First time I am using such a system thats why I need your reviews/comments.

I suspect there is simply not enough tension force in the chords to handle this type of situation. When I ran it, I got the following error message in the post-processing file:

"**ERROR-OUT-OF-BALANCE LOADS LARGER THAN INCREMENTAL LOADS AFTER ITERATION 153"

You said earlier that you used 5 kN pretension.....I doubt that would be enough with these kinds of spans. You may want to run hand calcs to confirm. (Or scrap this tension idea all together.)