Liquefaction Factor of Safety 1

[cbear1]

Hi All,

I am taking a poll to see what people are using for a factor of safety against liquefaction in California. We are in the Central Coast region, and it seems that many soils engineers use a factor of safety of about 1.1 to 1.3 for single family dwellings. The local jurisdiction requires a FS=1.5 for liquefaction. I have not been able to locate any FS requirements in the 2007 CBC at this time (please let me know if there is one..)

Is this overkill, or just a conservative requirement? What is the professional consensus for liqefaction FS?
Thanks in advance!

 

[howardoark]

Are you talking about slope failures caused by liquefaction?

I'm in the Bay Area and basically everything Holocene liquefies if it's saturated and not clay. We look at differential settlements (1/2 to 2/3s of the expected total displacement - though I've seen people use the difference in liquefaction settlements between adjacent borings as the differential settlement).

 

[cbear1]

In this particular case, there are no slope stability issues, just about 8 feet of sand with fairly low blow counts that can be considered below the ground water table. LiquefyPro was used to calculate the settlement, but a safety factor (factored into the CSS value)of only 1.1 was used, whereas the local jursidiction is requesting a 1.5 value which changes the results substantially.

I haven't found anything about this in the 2007 CBC and SP 117 suggests some values, but with no firm recommendations. Any help is appreciated.

 

[howardoark]

The Seed Queen Mary Paper (reprinted as a PEER report) has recommendations for calculating settlements due to liquefaction (if someone has a thought on how to adjust the settlements for silt content, I'd like to hear about it). The SCEC guide for liquefaction mitigation indicates that you should take half the total settlement as differential settlement (but doesn't provide guidance for the distance you should consider that differential displacement to occur).

Your factor of safety should really come from the return period of the earthquake you're considering (IMO). If you calculate the settlement from liquefaction due to the 475 year earthquake, take half of that and apply it across the column line distance (or center of mat to end of mat for slabs) and the structural engineer indicates he can live with that, I'd propose that to the local building official. If the structural engineer can't live with it, but indicates it won't cause the building to collapse and liquefaction mitigation costs more than the building would, then I'd argue it would be cheaper to replace the building than mitigate the settlements.

Or you could propose 50-year earthquake return period and 1.5 factor of safety.

It's the 00's baby, everything's negotiable.

 

[dgillette]

FS=1.5 seems a tad bit conservative for something with relatively minor consequences. It's not Oroville Dam or the TransAmerica Building or some major industrial facility that will cost millions to repair or release toxic chemicals.

That said, the simplified liquefaction triggering analysis has about a dozen sources of significant uncertainty, and H.B. Seed never considered the triggering curve to be a hard boundary. FS=1.1 may not provide much of a margin, but don't ask me what number is appropriate. If not for the code, you would look at the probabilities and the consequences. My guess is that they are low enough that an expensive foundation treatment could not be justified by any kind of risk calculation.

 

[sunny7uc]

Usually a FS of 1.1 is used to my knowledge.

 

[muuddfun]

I am in southern california, and everything that is below the groundwater table liquefies here, so I don't really look at the factor of safety, I just look at the settlement, bearing capacity, and surface mainifestation etc. And decide what to do about it. Anything on the borderline of liquefaction I think is considered to be liquefiable that is why you would use the 1.1 factor of safety. A factor of safety of 1.5 I think is high.

 

[BigH]

Can you explain how a factor of safety is given to "settlement" - that's what I seem to be hearing - I can understand if it is given to the capacity pressures but not applied directly to settlement.

 

[howardoark]

Factor of safety has little meaning for liquefaction since the susceptibility to liquefaction depends on the PGA and PGA depends on earthquake return period. So, the degree of conservatism depends on the return period you're looking at. If you have a 30 year period of interest, your liquefaction worries are much smaller than if you have a 200 year period of interest.

Factor of safety for liquefaciton is generally defined as the ratio of Cyclic Resistance Ratio (CRR) (mostly a function of density of the sand) divided by the Cyclic Stress Ratio (CSR) (mostly a function of acceleration and in-ditu stresses in the stratum of interest). That makes no sense to me, but that's how it's defined. Ray Seed and his co-workers provided various curves of probability of liquefaction - which while interesting, are also of little practical use.

Settlements are all based on lab tests on clean sands which you don't see much of in real life either. Generally, the numbers you get make you think that any stratum more than 1000 years old would be so dense it could never liquefy again - but that doesn't appear to be the case in real life.

Structures don't suffer (much) from just settlement. It's differential settlements that cause much of the damage and there isn't any sensible guidance for determining how much differential settlement occurs over what distance.

If you have a 20-foot thick liquefiable strata 30 feet down with a 20 foot cap of stiff clay over it, is that a problem? I don't personally think so, but my opinion isn't going to help you much with a building inspector who wants a factor of safety of 1.5 against liquefaction.

I think you have to use judgment. Which unfortunately allows clients to shop around for a geotechnical engineer who's judgment is for sale.

 

[BigH]

Thanks Howardoark - you got me to pull out Day's recently published Foundation Engineering Handbook - he covers this in section 13.4.3. Seen it before but just never put it to a "factor of safety". Isn't it amazing how building inspectors now drive engineering practice with code(s) as their bible(s)? (See some of the recent issues of Ontario's "Gazette" - the reporting of complaints against Ontario's Professional Engineers.)

 

[dgillette]

The real value of R. Seed's work is that they went back to scratch and reevaluated a lot of the things that went into H. Seed, Lee, and Idriss's charts and subsequent modifications - fines adjustment, Rd, CSR for each case history, K-sigma, magnitude (as proxy for number of cycles), etc. (Also added a lot more case histories.) This was long overdue, because the Seed-Lee-Idriss simplified method had more and more things added onto it without going back and making sure it all fit together, kinda like a car with the full J.C. Whitney catalog bolted on. For example, k-sigma came entirely, or almost entirely, from lab testing, and the magnitude weighting factor came from analysis of a very few records; R. Seed's work got them from the performance data. If you were to use R. Seed's 10% curve, it would accomplish about the same thing as Harry Seed's was intended to, which was NOT to be a deterministic boundary (this according to R. Seed and others who had talked to H. Seed about it when he was still alive). (Just don't mix parts from Seed-Lee-Idriss and Seed-Cetin-etc, like Rd from one and magnitude weighting factor from the other. They do not fit together, and you can't get an adapter kit from J.C. Whitney.)

I have no idea how the building code enforcers feel about all of this - a lot of people seem to view Seed-Lee-Idriss as deterministic. (For most of my work, I have the luxury of ignoring building codes, as long as I can get it past a board of heavyweight reviewers.)

 

[howardoark]

Day's recently published Foundation Engineering Handbook

For those of you who have Day's book, how do you feel about it? Personally, it was $120 that I wish I had back. Every time I pull it out looking for something, I never find it. So, not only did I waste the $120, it's wasted several hours of my time too.