Foundation Stability Factors of Safety 3

If it isn't a retaining wall there is no code basis for FOS is 1.5. Simply follow the code by checking all the load combinations against unity.

No current code basis.

What HD suggests is rational engineering although it may trigger a discussion with a reviewe

The 0.6D is essentially a 1.667 safety factor, rather than 1.5, is it not?

Yes, it is. The issue is that when you do 0.6D, and have a SF for overturning of 1.0 or 1.1 (which is actually 1.67 or higher), your bearing pressure is through the roof since the resultant load is barely on the footing. If I analyze it with 1.0D, I won't have those problems with the bearing pressures appearing way overstressed.

For example, with what I am analyzing now, at 0.6D + 0.6W, I have a SOF of 1.04, with max bearing pressure of 12.5 ksf (2 ksf allowable)
For 1.0D + 0.6W, SOF is 1.71 with max bearing pressure at 1.68 ksf.

This is an existing structure, so per 0.6D + 0.6W I would need to increase the size of that footing, which isn't necessary in reality.

I think current codes want you to consider the possibility that only 60% of the dead load could be there, not that you have a FOS = 1.5 against overturning. So you may have a bearing pressure issue.

Can you grab any other dead load to help you?

Since this is an existing footing, you may need to use engineering judgment to decide if this is really an issue, or just a code imposed issue.

DaveAtkins

I've said it before on this subject, I do ignore the 0.6D combinations for soil pressure calculations on isolated foundations resisting overturning. I do ensure that any 0.9D combinations have a OT FoS of at least 1.5. And I ensure the structure is still stable at 0.6D combinations.

These 0.6D combinations unnecessarily penalize soil pressure where the mass of the foundation is providing a large portion of overturning resistance. IMO, this is an oversight in the code.

And to be clear, my opinion above is for isolated foundations with significant overturning where the foundation mass is resisting a bulk of the overturning. I don't consider it applicable everywhere/elsewhere. Nor do I have any code justified defense for this approach besides the fact that it simply doesn't make sense to remove 40% of the DL and act like the foundation is actually behaving that way for soil pressure calcs.

DaveAtkins,

I can't think of a place where the codes specifically say to use ASD load combinations when checking soil stability, aside from retaining walls per IBC which says to not use the typical load combinations. It seems like a big gray area.

Geotech provides us with allowable soil pressures, so we should be using service loads and not ultimate loads.

Who's to say we shouldn't be using "Actual" loads which would be 1.0D or 0.9D rather than 0.6D for all soil stability calcs. And to clarify, I am referring to a structure where the majority of the dead load is in the footing and soil, as azcats mentioned.

I wouldn't think of is as only using x% of the dead load, my understanding is it is all about ensuring the same reliability between the LRFD and ASD combinations (see "Counteracting Structural Loads: Treatment in ASCE Standard 7-05" by Bruce Ellingwood and Yue Li)

The 0.6D + 0.6W combination checks global stability of the structure this is not limited to overturning checks only. As an example it is very easy to reach an overturning stability of 1.0 but still have sliding fail by a significant margin.

Celt83,

Thank you for including that reference. The examples used in the paper to back up the 0.6D all relate to superstructure or foundation anchorage (anchor rods) which are a different problem than foundation stability. I am not suggesting we use 1.0D + 0.6W when designing base plates or anchor rods. I am also not suggesting we neglect 0.9D + W when doing the strength design of the concrete foundation.

I agree about sliding, I am suggesting using a FoS of 1.67 for sliding, overturning and uplift when using 1.0D + 0.6W load combinations.

Thank you all for your input. I originally thought of using a FoS of 1.5 but believe 1.67 is more appropriate for 1.0D + (0.6W or 0.7E) load combinations.

I disagree that it is a different problem, the paper shows examples of various components of the stability load path. In my opinion the stability load path is only complete when resolved into the ground so if your calculations show the soil becomes significantly overstressed by the combination then to me that demonstrates a failure of global stability.

I think the source of the dead load/weight is something to consider here. If lots of dead load from super-structure components, finishes, partitions, equipment, etc. is being taken advantage of than I think the 0.6D combinations should be checked for stability type failures as well as bearing pressure. My thought is that the dead loads associated with the components described above tend to be overestimated to provide some extra cushion for gravity design.

If the dead load is purely weight of structure that is accurately accounted for than I think you could exercise some engineering judgement if needed.

I can't say I've seen any kind of commentary about this, just my 2-cents.

I think part of the problem is that a lot of the previous generation designed footings for 1.0D+1.0L and called it good. Now there is the question of whether an allowable bearing pressure is meant to limit settlement, in which case wind and seismic do not really matter, or stop soil failure, in which case they do.

HDStructural said:

I can't think of a place where the codes specifically say to use ASD load combinations when checking soil stability

Click to expand...

While it may not say use ASD, it has been the standard of care for as long as I can remember. But note, if you use 0.9D then you should be using 1.0W or 1.0E and not taking 0.6 or 0.7 as you will then need to use the LRFD combinations.

I think this is simple... 1.5 safety factor per IBC is for retaining walls, not site walls. 0.6D + 0.6W is a load combination in the code. So simply put, I don't see an argument for making your own load combination (using full dead) but then applying at 1.5 safety factor, just use what the code says of 1.0 SF with 0.6D.

The thing that gets me about foundation design in general is you can design a footing to work with ASD 0.6D + 0.6W, but then when you convert it to LRFD to do the concrete design using 0.9D + W, the entire footing could be unstable and you can't design it. I'm sure there are many foundations out there that work at 0.6D + 0.6W but then become unstable if checked for LRFD loads.

That's always frustrated me.

Back to the discussion at hand, the purpose of the 0.6D is for the factor of safety with respect to the 0.6W. The FoS between the 0.6D and the 0.9D load combos are a wash since we are increasing the FoS for the 0.9D case (they have the same effective FoS against overturning, sliding, etc.). The only thing that is affected is soil pressure as noted in my example above. For isolated footings, where the majority of the dead load is from the footing itself, 0.6D + 0.6W will never be an actual load that the soil ever sees. If this footing were to fail (assuming the soil had the capacity it should have), it cannot be due to bearing pressure being overstressed since 95% of the dead load will always be there, and the bearing pressure was adequate at that load combo. So for an existing foundation that is highly overstressed in bearing at 0.6D + 0.6W but within the acceptable FoS for the other checks, it doesn't make sense to increase the existing footing when it checks out okay for 0.9D + 0.6W.

Older engineers who taught me would sometimes increase the dead load of just the footing up to 0.9D but kept all superstructure dead load at 0.6D for foundation stability checks if they were constrained with foundation sizes. I get that it isn't by the book, but they don't lose any sleep at night for that, and I wouldn't either.

As others have said, it comes down to engineering judgement and is a case by case basis. I understand that my argument involves modifying a load combination and I respect all of your opinions.

Traditionally, a factor of safety of 1.5 has been provided for overturning, uplift, and sliding. The introduction of the 0.6D term essentially applies the factor of safety of 1.67 for you. The unintended consequence of using a 0.6D combination is that the bearing pressure gets hit with the 1.67 safety factor, on top of the safety factor of 2 to 3 that the geotechnical engineer already places upon it. The use of 0.6D generally leads to much larger footings when applied to the bearing pressure.

I'm baffled that the IBC resolves this issue for retaining wall footings but not building footings by allowing retaining wall footings to ignore 0.6D in exchange for a 1.5 safety factor on overturning and sliding. I assume this exception was intended to benefit geotechnical engineers who do retaining wall design and are unfamiliar with load factors, rather than being an intentional snub of building structural engineers.

Regarding this statement:

DaveAtkins said:

I think current codes want you to consider the possibility that only 60% of the dead load could be there, not that you have a FOS = 1.5 against overturning. So you may have a bearing pressure issue.

Click to expand...

Even if that were the case, it only makes sense for wind design, since wind lateral force is independent of seismic weight. If you only have 60% of the dead load, then you (usually) only have 60% of the seismic weight, and therefore 60% of the seismic lateral load. The code is treating wind and seismic inconsistently here.

This is clearly an oversight in the code that has doubtless increased the waste and cost of construction on numberless projects. I'm shocked it's been in the code this long. Anyone else paying close attention to what's governing their foundation sizes? 0.6D on the bearing pressure does it for me almost every time when a lateral system is involved. Does that sound right? You mean to tell me if I ratchet up the overturning moment on the footing it will fail in bearing pressure before it fails in overturning? I would suggest that is rarely the case.

Codes attempt to think for us. Engineering judgement is disappearing in favor of regulation. It's not my regular practice, but I wouldn't lose sleep over ignoring it (for bearing pressure calcs only). Foundations were designed for years without it. If you did get a local spike in pressure that caused an "failure", the soil would locally yield, and the soil pressure would redistribute over a larger area under the footing. If someone ever wanted to go toe to toe over this, a non-linear soil/foundation interaction model could demonstrate it's a non-issue.

This is like one of those old laws on the books that says you can't walk down the street with ice-cream in your back pocket.

JStephen said:

The 0.6D is essentially a 1.667 safety factor, rather than 1.5, is it not?

Click to expand...

Not quite.... I believe my previous in the thread listed below addressed this question pretty well. Essentially the 0.6D with a 1.0 safety factor is the same overturning stability as the old method (0.9D with a 1.5 safety factor).

thread507-450974

It's really more of a geotech issue than a code issue. They give you allowable bearing pressure generally controlled by long term settlement and the wind load is a very brief load.

I find it very odd that US codes use 0.6D. Meanwhile EU, AS, NZ and I'm sure plenty of other code use a much lower reduction factor of 0.9. The 0.6D goes against the basis of LFRD dead loads are not THAT variable and unknown. 0.6D implies that it the variability of dead loads is similar to that of live loads with is more than a little perverse. I realise that as many of you have pointed the likely purpose is to have a 1.5FOS on overturning etc and to alight LRFD to ASD. But if that is the goal then it should be achieved in other ways rather that false reducing dead loads.

DaveAtkins said:

I think current codes want you to consider the possibility that only 60% of the dead load could be there, not that you have a FOS = 1.5 against overturning. So you may have a bearing pressure issue.

Click to expand...

If that was the case then the it is quite odd that the variability of dead loads is so much greater in the US than the rest of the world.

DCBII said:

Even if that were the case, it only makes sense for wind design, since wind lateral force is independent of seismic weight. If you only have 60% of the dead load, then you (usually) only have 60% of the seismic weight, and therefore 60% of the seismic lateral load. The code is treating wind and seismic inconsistently here.

Click to expand...

Exactly. Reducing the stabilisation of dead loads while not reducing the seismic weight by the same about is illogical.

The 0.6 and 0.9 factors are de facto exactly the same. Because wind is reduced so much the end result is the same, just the reduction factor looks quite odd because ASD load don't play nice with actual predicted max wind loads.