Leg supports, thermal stresses, etc.
Leg supports, thermal stresses, etc.
(OP)
A design issue that comes up on occasion is a vessel supported on fairly short legs, with the vessel operating above ambient temperature.
To design the legs, I would usually assume the connection to the shell is "rigid", assume the legs are restrained horizontally at the foundation, and go from there, and that seems similar to the approach in Bednar.
The problem is, that with short heavy legs, if you use these same assumptions, then calculate radial deflections in the vessel from temperature and pressure, it will then give you very large radial forces in the legs. Attempting to beef up the legs is then counterproductive as it only increases the loads.
Meanwhile, if that leg can shift maybe 1/8" at the bottom, that load disappears entirely.
So how is that normally handled?
Is it reasonable to assume a base plate with bolts and grout is "fixed" for wind/seismic shear, but can shift enough to relieve thermal stresses?
Or assume the foundation can shift enough to relieve those forces?
Or use slotted base plates and bearing pad rather than a grouted column detail?
Or is this effect just neglected in the design?
A second question relating to the above-
In Bednar, he assumes allowable concrete stress of 0.25fc' under a baseplate, and I see similar recommendations elsewhere. Per the concrete codes, factored bearing can range up to 1.105fc' (including the A2/A1 factor) or with a load factor of 1.6 maximum, allowable bearing is about 0.69fc'. Is there some reason to use lower allowable stresses than what the concrete codes allow here?
To design the legs, I would usually assume the connection to the shell is "rigid", assume the legs are restrained horizontally at the foundation, and go from there, and that seems similar to the approach in Bednar.
The problem is, that with short heavy legs, if you use these same assumptions, then calculate radial deflections in the vessel from temperature and pressure, it will then give you very large radial forces in the legs. Attempting to beef up the legs is then counterproductive as it only increases the loads.
Meanwhile, if that leg can shift maybe 1/8" at the bottom, that load disappears entirely.
So how is that normally handled?
Is it reasonable to assume a base plate with bolts and grout is "fixed" for wind/seismic shear, but can shift enough to relieve thermal stresses?
Or assume the foundation can shift enough to relieve those forces?
Or use slotted base plates and bearing pad rather than a grouted column detail?
Or is this effect just neglected in the design?
A second question relating to the above-
In Bednar, he assumes allowable concrete stress of 0.25fc' under a baseplate, and I see similar recommendations elsewhere. Per the concrete codes, factored bearing can range up to 1.105fc' (including the A2/A1 factor) or with a load factor of 1.6 maximum, allowable bearing is about 0.69fc'. Is there some reason to use lower allowable stresses than what the concrete codes allow here?





RE: Leg supports, thermal stresses, etc.
Dirt, corrosion, rust, damage will jam the plates quickly.
RE: Leg supports, thermal stresses, etc.
You know these types of design problems and codes much better than I do, but is this a situation where you should put a big (in width and length, not thickness) repad on the top of each leg? The repad may not be needed for punching forces and the like, but so that the repad can flex and take the bending and displacement, without putting it all into the shell in such a concentrated fashion as might exist right at the top of the leg connected directly to the shell..
RE: Leg supports, thermal stresses, etc.
In that situation of thermal movement I've used slotted holes for the fixings and then calculated a preload for the fastener and combined it with a set of belleville washers so that the expansion can be allowed for without generating to much stress.
http://www.sealing.com/fileadmin/docs/Using_Bellvi...
The examples in the link above are for expansion in parallel to the bolt however provided you don't use to high a clamping load, the calculations will still work for you because your using the Belleville spring via the bolt to exert the preload as opposed to the bolt directly.