bracing thermal stresses

huangyhg

bracing thermal stresses
i'm currently working on a large rack in northern canada were temperatures can range from -45 to 35 celsius.  for the most part, the structure is designed to deflect so that it does not induce thermal stresses.  however, there needs to be an anchor bay to resist lateral loading.
most people i talk to say they would neglect thermal stresses in this braced bay and assume that the ductility of the bolted structure will allow for sufficient movement to absorb contraction and expansion.
the brace length is something like 11.5m between connection points.  given a construction temperature of 15 celcius or so, that corresponds to a maximum change of temperature of 15 - (-50) = 65 degrees celsius.
some quick math tells me that it's an contraction of 8.75 mm or so.  spec holes for 3/4 inch bolts are only oversized by 3mm.  so assuming both ends are centered in the hole, there's only 3mm of total sliding room.  
i'm not quite sure how to treat this.  there don't seem to be strict guidelines in the canadian code.  the nbc says you don't follow their temperature rules for various qualified codes, which include the steel code.  the steel code says to design for the load except if experience, math or engineering judgement say otherwise.  the more experienced people i've talked to don't seem to normally take it into account, and if i do, my bracing will be incredibly oversized in comparison to existing structures in the area that don't appear to have issues.  you also run into problems where increasing the bracing section sizes increases the temperature forces they induce, which increases sizes everywhere else as well to resist the bracing expansion/contraction forces.
also, a somewhat tangential issue that doesn't really apply to this project, what happens to a tension brace that's been heated up?
if a brace has been designed to act in tension and thermal expansion pushes a brace to buckling, it would seem that a large sudden deflection would happen if an applied lateral load had a direction shift.  wouldn't the structure have to travel extra distance to activate the brace in the other direction.
perhaps a k-truss could be effective.  the beam the truss   
structeit,
is the brace the only item that is subjected to the cold? i don't know what the structure looks like but if there is a beam/girder and columns either side of the braces (which i assume is the case)these would also contract albeit at different amounts.
let's make up an idealized example of a frame that has columns which are 15' tall, a beam that has a span of 35' and the x braces are 38' (11,6 m) long. also let's assume that the dt is 117f (similar to your 65c).
unrestrained by the brace, each column will shorten by approximately 0.137". we will assume all of this is taken by a reduction in height of the column because the column will not pull up the foundation. the beam will contract approximately 0.319" with assumption that the end of each column is pulled in toward the center half of this distance or 0.16".
the top of each column will be pulled down and to the center of the frame by approximately 0.21".
each brace wants to contract by approximately 0.348" (8,75 mm). like the column, only the end at the beam/column intersection contracts.
the net difference in length between what the brace wants to contract and the amount that the beam/column joint moves is 0.138" (3,5 mm)or about 40% of what would have first been thought.
again, this is idealized and i do not know if it applies to your geometry.
does this make any sense or have i missed something?
i know in pressure vessel design, many thermal stresses are considered to be a secondary, self-equilibrating stress and has a different (higher) allowable stress. i've never seen this concept applied to structural steel and do not know how it would be "codified" in a load combination.
patrick

structeit,
i am not sure i understand your concern and i have been designing steel structures in canada for many years.  why don't you provide us with a sketch of the structure in question so that we may consider it for ourselves?
ba
another issue to consider with steelwork thermal stresses especially is that they can heat up way more than max air temperatures under solar radiation. i have often wondered if anyone has any good info on what temps to use for thermal stresses.
dig1 has a good point that you can't look at the brace shortening in isolation. the beam and columns shorten too and that helps you.
i've done up a quick sketch.  unfortunately i'm on a site visit and don't have autocad or a scanner with me, so it's a pretty horrible mspaint abomination.
this anchor bay supports lateral loads from 5 bays to the right applied along the top and one very thin bay to the left applied at the mid-height   
try reversing the direction of your bottom truss   
that's not a bad thought.  the braces would end at the same node, so thermal lateral loads will mostly cancel out in the steel before they hit the foundation, plus the braces would terminate at an outside column where the tie beam will match some of the deflection.  
i'll take a closer look at it.  i haven't applied wind loading yet, so i'll have to see what that does to my lateral loads.  the change in bracing direction would mean that i have to transfer all shear through the centre anchorage.  i'll have to mess around with the foundation there and tie my anchorage into my reinforcement fairly well, since i have space issues that may not let me have a wide enough pile cap to develop the shear resistance for that anchor bolt group.  that's certainly do-able though.
braces shown in red in the attachment would distribute the horizontal reaction more evenly to the foundations and would have a reduced temperature effect.
  
ba