max. allowable displacemen

huangyhg

max. allowable displacement
hi!
be more specific. this type of connection for roofs is normally used for compensation of temperature induced deflection. if this is your case, you should compute the maximum of such a deflection and take it as a limit. if you can't give the structure as much sliding freedom, you must perform an analysis in which extra stresses due to tempreture is taken in to account.
for bridges, however, this type of connection is used to prevent fatigue due to the vibration. in this case you  should have no restraint against sliding.
theoretically, amount of longitudinal deflection would be a fraction of vertical deflection. in trusses, however, the vertical deflection is limited, so you shouldn't be normally worried about horizontal one, except you are dealing with a special problem.
if none of these problems you are facing, simply change the sliding type to a hinge one. it will reduce the size of the elements, but you should be aware of lateral reaction on the supports that could be as high as 1/10 of vertical reactions, or even greater.
good luck
visitor (visitor)29 jan 01 12:31
thanks for your response, gourile.  let me elaborate my question. the curved roof truss spans 130' and is supported on columns which are 45' high and fixed at the base.  if 2 pinned supports are used, there will be very large moment in the column, which is not desirable.  therefore, a sliding connection is considered at one support.  the truss is deflecting 4" vertically and 4" horizontally under dead plus live load.  temperature only gives additional horizontal deflection of 0.5".  do you think it is acceptable?  
you mention that "if a pinned support is used at two columns...", have you considered that the actual conditions after construction might create this condition?  in building construction, sliding connections are discouraged, particularly where uplift is a problem and due to the fact that sliding connections require maintenance.
consider that some moment will be in the columns due to the necessity of the connections and design accordingly.
hi!
for the conditions you described, i do recommend hinged connections for the truss because of the following reasons:
1- a curved shape loses its geometrical advantages for load bearing when simply connected. it can be proved by taking the bending moment of the midspan of such a curve and comparing it with a simply supported beam with the same span and load. it means that with this type of connection you go to the hastle of making everything curved for no value.
2- the horizontal deflection you mentioned would not be acceptable. if you proceed, you should think of a special kind of bearing to compensate. as ron mentioned, this type of connection will need maintanance which is not desirable for a building.
3- you have to make columns fixed to the foundation, whether you use a sliding connection or not. the problem arises when you are dealing with curved shapes. they will either useless or produce large horizontal reactions! i have two suggestions:
a)- with 130' (39m) span, you really don't need curved shapes. if you can, just avoid them!
b) if you have to deal with a curved shape, consider bracing the two exterior frames in the vertical plane, and connect them to the other frames by a system of roof bracing. then you should perform a 3d analysis so that the lateral reaction of the curved truss of each frame is transfered to the braced exterior frames by this system. with this trick' you will avoid large bending moments in the column base.
c) if none of the above is possible, try a double section column to reach high moment capacity with a low column weight and easy base connections.
good luck!
hello visitor,
primarily your curved roof structure will act in compression. this type of approach if in steel has a usual span range between 40 to 100m. well it does give u the long span but not suitable for concentrated loads.
as ron mentioned, the sliding joint in reality is a pin joint. the sliding/roller support serves  only for making a system determinate when  carrying out  calculations.
the deflections are high. is the bottom chord curved as well?
i would revisit the  span over depth  ratio for limiting the deflections and maintain  pin at one end and roller at the other. u could model this on staad effectively.
riz
visitor (visitor)30 jan 01 22:44
thanks again for your input, but let me clarify
1. ) sorry for the confusion, when i said '2 pinned support' i mean two hinges which are resisting both vertical and horizontal forces.
2.) i am considering a neoprene sliding joint because due to architectural constraint, i cannot get a column with the capacity to resist the moment from the horizontal force.
3.) there is no uplift problem.  the dl alone is higher than the suction from wind
4.) the connection is going to be exposed, so i guess, there shouldn't be any maintenance problem.
5.) i do have two exterior frames with diagonal bracings.  it does help cutting down the deflection.  however, i don't want to rely on them because the bracing elements are not going to be fire-proofed because of architectural reason, and i can't count on non-fire proofed elements to resist gravity load.
6.) the truss has both top and bottom chord curved.
it would be very much appreciated if you can give me further advice.
hi!
with your elucidations, i can just think of one solution. as you don't possess any type of load bearing i.e. a strong column or a reliable bracing system, your only choice is to make the curved truss so stiff that it deflects up to the allowable deflection range given by your design code. it would be a fraction of the span and you may use the allowable deflection given for beams with your own engineering judgment.
good luck.
since the top and bottom chords are curved, have you considered a cable stay to prevent lateral displacement?  if you can live with a cable spanning the columns, this would be an easy way to limit the lateral displacement, though you might have some architectural constraints that would prevent using a cable.
as you have probably seen in these forums, without as much information as you might have in dealing directly with the issue, we often throw out ideas just to see if one sticks!
good luck.
ron