Tubular Steel Structures-Welded Connection-Basics 4

IJR,
Your post is too vague for a comprehensive answer. Could
you explain your problem or project, so that specific answers
could be given.

For starters, (assuming you are designing an offshore structure),
you should be familiar with American Petroleum Institute's
API RP 2A (Recommended Practice for Planning, Designing &
Constructing Fixed Offshore Platforms) (Working Stress Design
or LRFD - Both are available). Minimum requirements for design
& detailing connections are given there.

Also, tell me what in the references are contradictory. May be,
it could be clarified

M. Hariharan

Yeah Hariharan: You are quite right. I didnt clearly state the problem. I am not new in metal construction, but apart from punching and overrall yielding of a connection, there is not much I know.

However I still believe I can use your help

Most references I have are recommendations, not codes, and apparently are written for guys like you, not beginners like me. And they dont tell why, they simply show graphs for use in design. This includes CIDECT publications and some papers found in AISC journals. I feel a bit uncomfortable with a ready made formula.

My structure is not offshore. It is a truss to support a roof. But it is long enough.

You have already answered a part of my question by suggesting a code API RP2, because what I was looking for is some minimum requirements for designing and detailing. If you would like to go on, please comment on those minimum.

Thanks for the concern

Respects

IJR

IJR,

Tubular trusses probably do not require too complicated
design calculations.

1. Analysis: Perform the analysis the same way as for any other
truss made of angles etc. (see below for member selection).

2. Member design: The interaction formulae are given in API RP
2A.If you had modelled the truss as a 'truss' (with pinned
ends) you don't need interaction formula. If the members have
K.l/r > 50, the bending effects would be very small, unless you
have member loading. A 'truss' modelling would be adequate.

There are advantages to keeping top and bottom chords of the
truss as I-sections. (Easy to connect tubular web members,
purlins etc)

3. Joint Design: If the truss chords are I-sections, you can weld
the tubular to the flange of the chord. Select the I-section to
have the requisite flange width. A pair of web stiffeners at the
connection point would be preferred, irrespective of design
requirements. The design procedure is complex and not worth
adopting for a roof truss.

If the truss chords are tubulars, the joint punching shear as
computed based on API RP 2A will determine adequacy of the
connection. The direct connection requires edge preparation
for the brace, this is available in API or AWS D1.1. At any joint,
the minimum distance as per API between two braces is 2 inches,
but for roof trusses, this may not be mandatory.

For roof trusses, gusset plates have also been used for
connections. The brace is slotted and inserted in the gusset plate,
and it is welded all around. A single gusset plate will receive
all the braces. This will permit centreline of all braces to meet
at a common workpoint. In such a case a channel section for
top and bottom chords could be advantageous.

The choice would depend on the availability & capability of
contractors in your area, and the relative economics.



l l < tubular > l l
l l l l
l ll l l l
l ll l l l
l ll l Gusset l l
\ ll / l l
\ ll/ =====
=== ll
ll ll I-beam
ll ll
ll Channel ll
ll =====
ll
===

Tried some figures, don't know if they will come out alright.

Hope this helps. Please ask if you need anything else.

Good Luck,

M. Hariharan

You have done it Hariharan. Respects

One final question

Assuming pin ended web members and continuous chords is an established procedure, because this can usually be assured if geometrical limits as specified by codes are satisfied.

But there are cases when a heavy brace has to be welded to an equally heavy chord. And there will be joint stiffness

Any rule of thumb to check out whether primary moments coming from joint stiffness should be considered or not?

Thanks once more and again.

IJR

IJR,

If heavy brace is welded to a heavy chord, the joint is a rigid
joint. There would be moments in the members. Conceptually
such moments should not be ignored. If the
external loading is only at the joints, these moments could be
small. If you are performing a computer analysis, it is simple to
model the members as frame members with rigid joints (i.e. no
member end release). The interaction formulae of AISC or API
can be used to verify the adequacy. If you are performing
manual computations (no computer) you may analyse the system
as a truss and limit the utilisation factors or unity check values
in axial tension/compression (actual stress/allowable stress) to
about 0.8-0.85. The balance would account for bending stress
due to joint rigidity.

Normally, tubular members are optimally selected if the K.l/r ratio
is about 80. For lightly loaded structures like roof trusses, you
may go for Kl/r of about 100. At a Kl/r ratio over 50, the bending
effects are quite small, and the above approximation would be
quite satisfactory. Of course, this excludes members which have
span loading. Normally, member sizes are dictated by Kl/r, and
available thickness.
If for some reason you are forced to use much
heavier sections, then a frame analysis may be required by the
authority approving your design. (If the section is much heavier
than required, it may alter the load path from normal truss
behaviour).

Hope this helps.

M. Hariharan

Thanks very very much Hariharan

If you visit here again, do you want to tell me how much temperature can exist in a truss member completely exposed to summer temperatures of around 80F.

Respects

IJR

In India, where the summer temperature can go over 100F, we
take that exposed steel can reach 60C = 140F. May be
you can reach 120F? Only a guess, no clue.

You must be having a sheeting over the truss. Then the
temperature would not go much beyond the temp. in shade.
If it is during construction stage, do you really need to worry?

Hariharan

If this one interests you Hariharan, I will go on a bit.

I have released my truss so that very little thermal induced stress develops in members. So actually I have nothing to worry about.

What I was trying to learn from you was the figure to be used in temperature loading. There seems to be no aggreement. As you mentioned above, temperature in steel can go well above ambient temperature in hot seasons. Do they get even colder in cold seasons. Is zero a reference temperature?

Not so important, but I have been arguing on this subject and going in circles and that makes me a bit unhappy.

It is nice to see you also visit AISC forum.

regards
IJR

IJR,

We have not considered thermal stresses for offshore
structures here in India, primarily because the structures are
shielded from direct solar radiation, and the variation in
ambient temperature is rather small. Exposed truss structures
such as bridges are generally &quot;statically determinate&quot;, if you
ignore the rigid moment connections. They are
designed for cyclonic wind conditions, and when the cyclone
(same as hurricane in US) blows, the temperature is quite close
to normal.

I did not mention earlier, you should have one base joint a
sliding connection, so as not to introduce thermal stresses in the
truss. The size of the slot for bolt hole would need an indication
of the likely variation in the temperature.

An exposed steel has a much higher temperature than ambient
in daytime on account of solar radiation. However, there is no
mechanism to lower the temperature below minimum. I am not
sure if wind chill factor is applicable to metals! Probably some
slight reduction in temperature could be expected. I would
consider the temperature for any evaluation to be the
temperature at the time of fabrication / erection.

Incidentally, we designed a structure for a Norwegian project.
Design basis & criteria were provided by the client.
The environmental conditions included 'ice-condition', wherein
each member was assumed to be coated with ice of a certain
thickness, and storm wind acting. Even this condition did not
incorporate any thermal stresses! We had different conditions of
fire and increase in temperature of steel, but the increase was
considered only for determination of reduced elastic modulus
and yield stress, not for thermal stresses. Considering how
detailed the computations were, I can only surmise that the
basic reason would have been that the stresses would not have
been very large. Honestly, I had not given a thought to it at that
time. We used limit state design, and that distorts the values
somewhat due to the different load factors. May be we would do
some number crunching to find out if thermal stresses are
indeed small.

Nice discussions!

M. Hariharan

Can not thank you enough Hariharan. I can however wish you all the best and perfect healthy life

regs
IJR