Design of Saddle for Horizontal Vessel 4

[AK8824]

Compress uses a Friction Coefficient of 0.45 for steel-to-concrete contact. This results in larger resultant forces and the anchor bolt size. Can the Gurus advise any better way? using a Teflon pad or SS plate seems one of those possibilities.

 

[SnTMan]

AK8824,

Not sure what problem you're trying to solve (also not familiar w/ Compress), but assuming you are trying to size the anchor bolts against such things as horizontal seismic loading or bundle pull of a heat exchanger:

Friction is to your benefit as the friction force absorbs part of the load that would be otherwise transferred to the anchor bolt. It seems to me that a similar situation would exist if you were, say, looking at bearing of the bolt against the concrete foundation.

In designing supports for horizontal exchangers it is common to assume zero friction between support and foundation. This tends to oversize things to some extent, but unless there are other considerations, the extra ruggedness is worth the small cost difference.

Hope this is helpful.

Mike

 

[unclesyd]

We use a very heavy asphaltic felt under all horizontal vessels Al, CS, and SS. We have all concrete saddles. The felt material is about 1 in thick and as wide as the saddle. It is quite similar to material used in concrete roadway expansion joints only a heavier weight.

 

[jte]

AK and SnTman-

In my past life we used a friction factor of 0 for sizing the anchor bolts on the non-sliding side for shear during a seismic event (presuming there could be some vertical component which would thus eliminate any friction on the sliding side). For evaluating thermal expansion loads on the saddle webs/flanges/ribs we used a friction factor of 0.3 for steel baseplate on a concrete foundation/pier or 0.1 for steel on teflon. I'm not sure where the 0.3 or AK's 0.45 came from, but apparently there are differences of opinion out there... When in doubt, check with the client for their opinion.

jt

 

[AK8824]

The context is design of Horizontal Pressure Vessels of larger size, say 4000 mm dia and more. Compress is one of the well known design software, which uses ASME code as the basis.

Usually, one of the two saddles is anchored, the other one remaining free to slide. Depending on the weight and the coefficient of friction, the fixed side anchor bolt will experience a shear, and it must be designed to withstand that.

Teflon is one such material that can reduce the coefficient to between 0.1 and 0.15.

 

[stanier]

The slide bearing could be mounted on "rollers" and your friction would be very low. There would be a need for restraints for the earthquake case.

Geoffrey D Stone FIMechE C.Eng;FIEust CP Eng

http://www.waterhammer.bigblog.com.au

 

[TomBarsh]

'COMPRESS' assumes that one saddle is fixed and that the other saddle is free to slide in the longitudinal direction. Forces acting on the vessel in a horizontal direction are applied to the fixed saddle and the "free" saddle is assumed to slide as necessary. Or more to the point, the free saddle cannot sustain any force in the direction of the vessel axis.

But the assumption of a saddle base plate that is free to slide is relative. As we all know, if we push on the kitchen table a certain amount of force is required before it begins to move across the floor. Once the table has started sliding less force is required to keep it moving. These effects are the result of friction. When the table is stationary the "static" coefficient of friction is effective in resisting movement of the table. Once the applied force exceeds the "breakaway" force the table begins to slide because the applied force has overcome the maximum resisting force possible with the static coefficient of friction. Friction still exists once the table is moving but the coefficient of friction is smaller and is referred to as the "dynamic" coefficient of friction.

The coefficient of friction used by COMPRESS is the static version. This value must be entered by the designer based on the actual conditions of construction. The default value, 0.45, is relatively high and would apply for a rough finished steel base plate on concrete or grout. The high value is used because it leads to a conservative analysis (as will be seen later).

Why is the frictional force at the "free" saddle even considered? Consider a vessel subject to a high temperature process, or even to very hot ambient temperatures (sunny day in Saudi Arabia). The vessel length will "grow" due to thermal expansion. If this growth is resisted by mechanical constraint then a force results. The resulting force is applied to both saddles. The "free" saddle is subject to this force but it cannot slide or move until the breakaway force is exceeded, at that point the force is relieved by movement of the saddle. Thus the maximum possible force that the free saddle is subject to is that resulting from the coefficient of static friction. The force opposes the direction of movement and is the product of the normal force (in this case, the weight) and the coefficient of friction. Of course, the actual force due to thermal expansion may be less than the maximum possible force as considered by COMPRESS.

The anchor bolts at the fixed saddle must resist the wind force, seismic force, and the force due to thermal growth. COMPRESS applies the largest of these three forces to the anchor bolts at the fixed saddle. (Perhaps this should be the larger of wind and seismic, combined with thermal growth.)

The "free" saddle can be made more "free" by providing a base that has less friction. This can result in eliminating thermal growth as the controlling condition and possibly may allow use of smaller anchor bolts. Less friction can be accomplished by any number of means: a metal wear plate on which the saddle base plate rides, lubricated surfaces, elastomeric bearings of some sort, or teflon slide bearings like Fluorogold bearings. Fluorogold isn't too expensive and is used a lot for pipe supports to allow movement of the piping; it has a very low coefficient of friction (offhand I do not remember the exact value). Also, don't overtighten the nuts on the anchor bolts at the sliding saddle!

One eng-tips member posted awhile back about some very large horizontal vessels in Saudi Arabia. These things were around 20' dia x 200' long on two saddles. I think that consequently there would be high loads (weight) acting on the saddles along with a high degree of thermal expansion. I wonder how they detailed the sliding saddles for those vessels. That idea of a slide bearing with a roller (just like a highway bridge) would likely work very well there.

Tom Barsh
COMPRESS technical support (but responding on my own behalf)

 

[panduru]

Thank you, TomBarsh, for the excellent post.

I would like to see you more often in the forum.

 

[LSThill]

Design of Saddle for Horizontal Vessel Friction Loads:
The following friction coefficients shall be applied in the determination of friction forces:
Concrete to Soil………………………….. U = 0.50
Concrete to Polyethylene Sheet……U = 0.50
Steel to Concrete………………………… U = 0.45
Steel to Steel…………………………….. U = 0.40
Steel to Teflon…………………………… U = 0.10
PTFE on PTFE.................. U = 0.10
(PTFE: Poly-Tetra-Fluoro-Ethylene) ?

 

[panduru]

@lsthill:

> Yes. PTFE = Poly-Tetra-Fluoro-Ethylene. Btw "Teflon" I believe is a trade name and refers to the same thing as PTFE.

> Steel - Teflon @ 0.10 - I am a bit sceptical about this.

> Source? Can you please mention the source of the above information, thanks.

 

[AK8824]

Thanks for your contribution on the topic. Tom has thrown a good amont of light on the topic, which benefited me.

 

[Shmulik]

panduru:

Quite detailed info you may find at the following:

http://www.roymech.co.uk/Useful_Tables/Tribology/co_of_frict.htm

 

[panduru]

Thanks Shmulik, for the excellent link.