Liquefaction Induced Settlement at Ground Surface

[dirt-nerd]

I'm not interested in calculating the liquefaction induced settlement in each liquefiable layer. My interest is in predicting/estimating how much that cumulative settlement will be manifested at the surface. Intuition tells me that some of the non-liquefiable material (or crust) will bridge over the settlement or absorb some of the settlement in increased volume in said layer. That would result in less settlement at the ground surface (or BOF elevation) than the cumulative amount in each liquefiable layer. However, I'm struggling with reasonably estimating this.

Ishihara (1985) seems to intend for “surface manifestation” to refer to sand boils, ground fissures, etc. and for “surface manifestation” to not refer to ground surface settlement. Essentially, no surface manifestation =/= no ground surface settlement.

Rateria (2021) expanded on Ishihara’s model with thousands of data points from recent New Zealand earthquakes; this paper states:

…placed emphasis on compiling case histories from free-field level-ground sites, with the occurrence and severity of liquefaction defined primarily by liquefaction ejecta and ground cracking. Sites with other indications of liquefaction, such as foundation settlements or evidence from ground motions, were not considered. The definition of ‘‘surface manifestation” adopted herein is thus generally consistent with that used by Ishihara (1985) to develop the H1–H2 model.​

Here, Rateria seems to reiterate that the H1-H2 model is not applicable to ground surface settlement but rather the H1-H2 model is intended to predict surface manifestation (sand boils, ground fissures, ejecta, etc.). However, later in the document, it states:

It should be emphasized that ‘‘surface manifestation” – the prediction target of the analyses presented herein – refers to free field surface ejecta, settlement, and cracking on ground that is generally level, consistent with the definition adopted by Ishihara (1985).​

The way geotechnical engineers use the H1-H2 seems to be inconsistent. I've seen some not use this method to reduce the settlement at ground surface and seen others use this method to reduce the settlement at the ground surface.

How do people employ the H1-H2 model and how do you estimate the reduction in settlement at the ground surface from the cumulative settlement from each liquefiable layer? Interpret the H1-H2 model as exclude ground surface settlement if the H1-H2 model shows no surface manifestation? Or that you cannot use the H1-H2 model to predict ground surface settlement? Are there other methods that are better or clear standard of practice? I'm familiar with a Bray (2017) method to estimate liquefaction-induced building settlement but I've never seen it been used in practice.

Thank you in advance.

 

[geotechguy1]

While not a direct answer to your question, there are geotechnical modules published by the New Zealand Geotechnical Society which contain alot of pragmatic information for application of these methods in a country where there are alot of liquefiable soils.

https://www.engineeringnz.org/news-insights/revised-geotechnical-modules-released/

and the MBIE liquefaction document:

https://www.building.govt.nz/assets...ucture/planning-engineering-liquefaction.pdf.

And the American national research council document:

https://nap.nationalacademies.org/c...nduced-soil-liquefaction-and-its-consequences

>The way geotechnical engineers use the H1-H2 seems to be inconsistent. I've seen some not use this method to reduce the settlement at ground surface and seen others use this method to reduce the settlement at the ground surface.

My take on it is that basically in the H1-H2 method you are hoping that if you have a thick enough crust and no major shear induced building settlement that the crust will sort of smooth out differences in differential settlement and the ground ends up settling a similar amount across the area with that crust - which from an engineering perspective may be OK. There is a good photo in Johnathan Bray's presentation or research somewhere that shows a building in Japan where one building is on a shallow foundation and settles 500mm more than the free field due to shear-induced settlement from the building, the adjacent building is founded on piles and settles 500mm less than the free field settlement of the adjacent ground. Which is better performance? Both probably had their service connections sheared off and both are now half a meter different than the surrounding ground.