Anchor Rods for Thermal Growth of Horizontal Vessels

[JAX91]

My company does a lot of foundation designs for horizontal vessels, and I am looking for some insight on anchor rod design for these vessels. The vessels typically sit on two saddles, one on each end. One of the saddles is slotted to allow for thermal growth of the vessel. We normally put a slide plate under the sliding saddle, and use grout under the fixed saddle. We design the foundations for a thermal growth load equivalent to the coefficient of static friction between the slide plate and the saddle times the weight on the saddle. In the past, I have also designed the anchor rods for this same load; however, I am now thinking this may be overly conservative. The way the system is designed, the compressive force in the vessel increases as it heats up. It continues to increase until it overcomes the static friction at the sliding plate end, and then the vessel slides, releasing the load. Ultimately, the anchors on the sliding end never see a load from thermal growth because of the slotted hole. Because the load is released on the sliding end, the friction on the fixed end is never overcome. If this friction force is not exceeded, the vessel cannot slide to engage the anchors. Does that mean the thermal expansion never places load into any of the anchors? Essentially, if the vessel was set on the same grout/slide plate configuration with no anchors and no other transient loads, I would anticipate the same exact result. The fixed end would never move, and the sliding end would only move longitudinally with the thermal growth of the vessel. Does this sound correct?

 

[racookpe1978]

In theory, yes. You will often see rollers rather than easy-to-rust, easy-to-foul or jam sliding plates if the load can be made small enough.

Most often, it is more than just friction holding the fixed end - the bolts and bolt holes themselves are designed against shear and for just enough "slop" for construction tolerances.

 

[JAX91]

Lately we have been specifying slide plates made with Tivar Dry Slide. It is a UHMW plastic plate with friction and compressive strengths comparable to Teflon. Because it is plastic, rust and other maintenance issues are not an issue.

 

[racookpe1978]

Keep them clean! A physical jam will stop movement, even if not rusted or galled between the moving surfaces.

 

[JoshPlumSE]

The anchored end is NOT slotted is it? If not, then those anchors do see the full friction load. Maybe you can say they don't if you are using a slotted plate on the anchored side as well. But, I don't think that would work for wind or seismic at all.

 

[JAX91]

The anchored end is not slotted; however, the holes are slightly over sized for construction tolerances (typically 1/4" oversized). I am looking at the situation more like a cantilevered retaining wall. In order to develop your passive pressure, the wall has to move a little. In this case, in order to develop resistance from your anchor rod, the anchored saddle would have to move a little bit. If it doesn't move, the load is resisted by static friction between the saddle and the grout, not by the anchors. Because of the slide plate system, the only way the anchored end could move is if the slotted end was not slotted enough and was not allowed to slide (or if it binds up as racooke mentioned). I agree, the wind and seismic would still need to be checked. But in reality, if these loads were not great enough to overcome the friction, they would have the same effect. Any thoughts?

 

[JoshPlumSE]

In theory, I think you are right. If all goes according to plan, those rods won't see the full force. I just flipped through an design manual on the subject (written in the 90's) and in it my previous company acknowledged exactly what you suggest. That you will have to overcome friction before those bolts see this thermal force.

Of course, if you have a wind or seismic event that could enough slip. And, the reason for 1/4" oversized holes is that rods are frequently misplaced slightly. Or, you could get some corrosion or something that jams the slide plate (as RACooke suggests). For those reasons, I tend to think it is reasonable to design these rods as if they do see the force. But, I can understand the argument for not doing so.

ASCE has a document on the design of anchor bolts for petrochemical facilities. I don't have a copy, so I don't know if they directly address this subject. But, it looks like section 3.6 (based on the table of contents) might address it. It's probably worth obtaining if you want a more definitive answer from a source you can reference. Give yourself a bit more ammunition if a client or plan checker or peer review ever question the decision.

 

[JAX91]

I was able to track down a copy of the ASCE Anchorage Design book in my office. The first sentence of Section 3.6 states, "Anchors need not be designed for shear if it can be shown that the factored shear loads are transmitted through frictional resistance developed between the bottom of the base plate and grout at the top of the concrete foundation." The section also gives a coefficient of friction between steel and grout of 0.55. The slide plate I am using has a coefficient of friction of 0.16. Once the slide plate friction capacity is exceeded, the force is dissipated through the movement of the vessel. Assuming the slide plate never “binds”, the thermal load will never place a load on the anchors. In my case, if the slide plate binds, the load transferred into the anchors at a temperature difference of 20 degrees F is over 1.5 million pounds. I would never be able to reasonably design for the load, so I have to make the assumption that the slide plate would never bind. The only other case I could check would be partial binding of the slide plate where the coefficient of friction may be increased. If I design for a higher coefficient of friction, there wouldn't really be a reason to provide the slide plate.