Foundation Bolt Design

[dozer]

Has anyone ever seen an article called "Foundation Bolt Design" in the Dec. 24, 1980 (that's right, 1980) of "Plant Engineering"? It was recently presented to me as the way to design anchor bolts in a circular pattern. It does not agree with the 4M/dN - W/N for bolt tension found in most references.

I personally think the article is just plain wrong. It appears the equations are derived from assuming the bolt pattern pivots about the edge of the bolt circle. In other words, every bolt is in tension!

I've pretty much written it off but I was wondering if any of you history buffs know if the article was ever formerly (or informally) rebutted?

 

[JStephen]

I've run across that idea in specifications. I think it relies on faulty reasoning, but then again, I haven't actually seen a technical discussion of the idea, just the specification requiring it.

As I recall hearing the idea, people would take a cross section through the concrete immediately below the surface, and then treat it as a composite steel-concrete section as is commonly done in reinforced concrete design. The problem with this is that if you take a cross section immediately above the surface, you'll get a different load distribution, but the tension in a particular bolt can't vary over that short of a span. Both these assumptions are based on beam distribution, which is derived for long cross sections, so neither is strictly applicable to the case at hand.

I noticed that in one of the latest tank standards, they give the load on the anchor bolt as the uplift in the shell times the spacing of the bolts. I think this is perhaps specifically intended to direct people away from the composite-section approach.

 

[kbarnett]

If the base plate can be considered as being infinitely rigid, then I could see how an argument could be made for all bolts being in tension. You would run into bearing issues if the base plate were able to perform like that. That also assumes that the base plate bears on concrete and is not suspended by the anchor bolts, as is the case with some designs where the columns are leveled with the protruding threads of the anchor bolts.

Sorry, I can't help you with locating the article. I know of some that have to do with base plate design if you're interested.

 

[steve1]

I think that the nature of the problem is the determination of the "correct" location of the neutral axis on the plane over which the moment acts.

Each of the techniques described on the previous posts calculate the bolt tension based on the assumptions used to develop the equations. Assuming no mathematical errors, each will produce correct results within the context of the inherent assumptions. The most conservative results would be obtained with the equation that produced the least distance between the centers of gravity of the tension and compression zones.

In a practical sense I believe that any of the equations could be used. I would go, in the absense of any contrary specification, with the system I was most comfortable with and felt I could justify if called upon to do so.

 

[dozer]

I've got the article. I was wondering if anyone else has seen it. Sorry, I disagree with the statement that any of the equations could be used. An engineers job is to predict how things will behave in the real world or at least make an estimate that will provide conservative results without breaking the piggy bank.

The method presented in this article produces what I believe are unacceptably uncoservative results. I did a FEA of the base plate and got a maximum bolt tension of about 17 kips. The method proposed in the article got 8 point something kips. The hand method I used gave reasonable agreement with the FEA.