deflection improvement for the existing shed

huangyhg

deflection improvement for the existing shed
hi fellows
a urgent question. i designed a factory shed last year, and they are building it now. the rafter is 310ub slope from one end to another, span 20m, with 3m haunch both end. the column is 360ub, high 3.6m at on end, and 5.2m at the other. the rafter rigid connected to the column. there is a bolted splice at the middle of the rafter. (top and bottom flange, web bolts with plate).
now my problem is, in my calculation, the rafter is deflect 43mm under dead load, but now during constrution, the beam is deflect 30mm only under self weight, so i guess it will deflect more than 80mm under deal load (included purlin and roof sheeting). the owner is not happy about this deflection. do your guys have any idea to improved the deflection without major change of the design? i think the main reason for this happen is because the bolt splice, in my member design, i consider it as a single beam, but because 310ub not available for 20m long, so i designed a bolt splice at the middle, may be the bolt hole and the bolts create some rotation.
thank you for your time.
powerranger
i'm not sure from your post.  did you use a splice on the rafter that connected the top flanges and the bottom flanges of the rafter?  the web bolts won't transfer moment, until it is "ecessively" deflected.  if you did connect the top flanges, go out and look at it.  assuming you used bolted top and bottom flange connections to transfer moment, are the bolt holes in the top and bottom flange oversized?  did they allow the bolts to move within the holes, creating the deflection?
also, you mentioned that you rigidly connected the rafter to the column.  when you calculated deflection, did you consider that the column will rotate?
assuming the splice is causing the deflection, what about jacking the beam up and welding the splice?
is the splice designed with hsfg bolts or as a simple connection? are the mating surfaces of the joint painted? i expect that the joint was not bolted together correctly and that the bolts have moved within the holes.
i have had a similar case previously and we jacked to correct the deflection, then we loosened the bolts and jacked another 30mm. following retightening the bolts we welded around the splice plates.
the question is whether to wait until all the deadload has been applied and then to carry out the correction. this may provide a more accurate final result.
i did a couple of quick checks on using the values you provided, making some assumptions where necessary.
for a 305x102x25ub (which is admittedly the smallest) the maximum deflection considering a simply supported   
thanks guys
hi ussuri, the second moment of area of my ub is 86.4x10^6 mm^4. if you consider the span between haunch, it will be only 14m, which will give a 45 span to depth ratio.
hi tomfh, zambo, jacking the rafter up and re-treat the bolting splice can help, but need to jack up not just the mid span, (because the rafter will sag between the column and the jacking point, so at least 2 more jacking point in between is required).
but anyway, the owner agree to "suffer" this happening. but one thing he mentioned, and make me feel that not sure right or wrong. he said, if we know the rafter will sag at the design stage, why did we not design the rafter with greater pitch, than, when it sag, it will come down to the pitch he want. he get his point, but one thing make me not sure is, later, does the drawing from the architect, shows the pitch of th rafter, should i consider it as the final pitch in my design, or the initial pitch (which used in the model).
thank you.
is the design still adequate with the splice at the middle? the bending moment you've lost at the splice joint will redistribute to the corners.
hello pr
comments on a number of points;
-"the beam is deflect 30mm only under self weight, so i guess it will deflect more than 80mm under dead load"; this is not necessarily true; fabrication/bolting/joint slip could all have contributed to the initial deflection and possibly won't contribute (much) to additional deflection.
-"is the splice designed with hsfg bolts or as a simple connection? are the mating surfaces of the joint painted? "; good questions, bolts must be grade 8.8/tf and properly tensioned, and either no paint, or compatible primer, under the splice plates.
-"i have had a similar case previously and we jacked to correct the deflection, then we loosened the bolts and jacked another 30mm. following retightening the bolts we welded around the splice plates."; good advice
-" is it possible the frame is just not stiff enough?"; i would ask the same; my rule of thumb is 19-20mm per metre span, 15mm is low.(i'm not saying it isn't correct, however it is at the light end)
-" but need to jack up not just the mid span"; i don't see why, you are only correcting (or preferably over-correcting) the rotation across the splice.
-"why did we not design the rafter with greater pitch"; yes, it is common for long span beams such as this to be designed with a pre-camber or pre-set to counteract dead load deflection. for lower pitch roofs dead load defl'n such as this could cause ponding and leaks.