How do I average N values from an SPT test with multiple boring logs for a specific depth? 1

[AdonisDNA]

I'm working on an assignment and have been given a soil profile with 3 boring logs. Each boring log has N values at different depths using the SPT test. The N values have been taken from 3 different soil layers.

The boring logs reach a depth of 20 m and I'm interested in the N value of a sand layer that is only 10 meters deep. I've collected all the N values and depths for each boring.

This is my current method:
[ol 1]
[li]Average the N values for each boring per soil layer[/li]
[li]Average the average N values for each soil layer[/li]
[li]Take the average N value for my sandy layer only[/li]
[/ol]

Is this methodology correct?

 

[GeoPaveTraffic]

I would omit step 2 and just average the values for each soil layer.

Depending on how variable the data, I might use a value somewhat lower than the average to account for the softer/less dense material.

Also, if one boring has consistently lower N values for a given layer, then you might want to use just that boring's values in your analysis for that area.

It all comes down to judgement and trying to interpret the site based on the data you have. Remember that nature is very unforgiving and a weak zone or layer will almost always control you design.

Mike Lambert

 

[Ron]

Each boring is distinct. Do not average N values across borings for analytical purposes. OK to do if you are providing general information.
As for layers...if you have visual classifications of a particular layer, you can average the N values in that layer (as GPT noted)

 

[AdonisDNA]

After taking the average of each boring for each layer, I decided to take the lower number between the boring averages in order to control for the worst. In a real world example, I think this would be a good way to go about it.

I looked everywhere for an answer to this online and couldn't find it. I guess it pays to join a discussion and just ask.

Thanks again

 

[BigH]

Okay - so you have the "N" values - what are you going to do with them?

We do not know what the purpose of this exercise is. I might suggest you read Peck Hanson and Thornburn's Chapter 19. He has an illustrative example on how to "average" the N values and if corrections are or are not needed (this is for spread footings on sand). Terzaghi, Peck and Mesri (article 50.2.5) also discusses the use of the N value in design of footings on sand (more elaborate than it used to be - see section 50.2.8 as well). If you are using the "N" value for deep foundations, you might wish to see TP&M article 52.3.7. One can also check out Kurian's book (Design of Foundation Systems), Chapter 4.1.3 and .4 if for shallow foundations on sand (dubious at 10 m deep - although there might be an excavation that takes the footings that deep).

There is information available and I suggest that one should develop a library of top quality texts (Terzaghi, Peck & Thornburn, Tomlinson (if using 6th edition or later, he uses Eurocode a lot - his older (3rd edition, for example) are better in my view for the "beginners"), Peck,Hanson and Thornburn, Tschebotarioff, etc.). One can also find the Poulos paper at the Istanbul ISSMFE conference (2001, I believe) - good paper.

For any real help, you should advise nature of project, a "Model" of the borehole stratigraphy, deep or shallow foundations, exavation, etc.

 

[fattdad]

BigH post is correct. Without knowing the intent, it's hard to comment.

My typical approach is to average the N-values from any given boring for the layer in question. If you have three borings, you'll have three averages. Take the lowest.

Items to consider: Were you in the field when the borings were performed, was it an automatic hammer or a safety hammer, do you know the drilling company, is ground water a factor, was it mud rotary or auger drilling, etc. Now I'm not saying you make calculated adjustments for each of these items, I'm just saying if you are dealing with an unknown driller, there is no professional logging, you are in sand and drilling below the water table, you may have bogus N-values from the onset.

It's the GIGO factor.

f-d

¡papá gordo ain’t no madre flaca!

 

[Mccoy]

The boring logs reach a depth of 20 m and I'm interested in the N value of a sand layer that is only 10 meters deep. I've collected all the N values and depths for each boring.

This is my current method:
1.Average the N values for each boring per soil layer
2.Average the average N values for each soil layer
3.Take the average N value for my sandy layer only
Is this methodology correct?

I agree that the OP did not specify the intent, but, given the depth of interest, let's hypotethically assume it to be shallow foundations, maybe a building with an excavation.

The correct method is the following in my opinion:

1) Average the N values for the sand layers in each boring
2) Run a statistical test comparing averages (like a simple ANOVA) and see if they belong to the same data population
3) If yes the layer is laterally homogeneous, then, pool all the data together
4) If not, then try and single out the datasets-borings which do not belong to the same population and pool all other data together, since the layer is laterally dishomogeneous.

The above implies that there is not a significant NSPT trend inside the sand layer, otherwise it should be broken down into vertically homogensous sublayers or maybe we should follow some specific statistics which considers the trend itself.

Please note, if the above sounds arduos to you, you can proceed by the visual method plotting the N counts and overlapping them to see the differences. Actually, I follow the visual method first thing myself, nothing beats a trained engineers' eye!

 

[Mccoy]

Adonis, maybe now we know that those SPTs were done on the tank site!