Effective strees analyses in slope stability applications 2

[georam]

I would like to get some comments or input on the selection of c’ and phi’ for cohesive soil<br>
properties in slope stability analyses using effective stress analyses.<br>
<br>
Let us say, that based on a considerable amount of direct shear tests, it was found that the soil has <br>
the following properties (same data, by regression analysis) :<br>
<br>
1. Regression - not forcing result to c’=phi’=0 <br>
c’= 20 kPa<br>
phi’= 18 deg.<br>
<br>
2. Regression - forcing results to c’=phi’=0<br>
c’=0 (forced)<br>
Phi’= 22 deg.<br>
<br>
In the steady state seepage analysis (long term), we usually are reluctant to include any c’ component in the effective stress analyses. Adding a small amount of c’ in the stability analyses would affect the results considerably. However, if we choose the properties based on No. 1 above and dropping the c’ to 0, then we “internally” penalize (over conservative) in just using phi’=18 deg.<br>
<br>
What about using the properties as shown on No.2 above ? (ie., we accept that c’=0, but another rerun on regression analysis will be carried out, forcing the data to include<br>
c’=phi’=0. This way, we will use the strengths of c’= 0 and phi’= 22 deg.<br>
<br>
Of course, if we choose No. 1 above and minimum F.O.S =1.5, it will be too conservative. What about the option of using data from No.2 ? <br>
<br>
Any comments ?<br>
<br>
PS.<br>
Some people tend to analyze using concept of No. 1 above, but then they add a little cohesion in the analyses, such as using minimal value of c’= 4 kPa (0.5 psi), instead of zero. So, which is the right way for effective stress analyses ? <br>

 

[Ginger]

Dear Georam<br>
<br>
I'm not sure that there is a &quot;right way&quot; to do the analysis based solely on the estimated parameters. From speaking with other engineers the best advice I have received is to initially look around the area for existing slopes in similar materials (ie railway or road cuttings etc). This will give you an instant verification of what slopes are achievable in the drained condition. If you have undertaken a series of on site tests then this will give you your total stress parameters. Rather than simply relying on regression analyses of these results alone I would also be tempted to find parameters successfully adopted for other engineering projects in these materials. <br>
<br>
The final parameters you choose should ideally then be based on a range of parameters for c and phi and you can check the sensitivity of your results against the risk of slope failure. Whilst this may seem like a lot of work, most software these days takes the grind out of the calculation but you get a good appreciation of the sensitivity of your design and the choice of parameters becomes almost secondary to your engineering judgement. <br>
<br>
Regards<br>
<br>
<br>
<p>Andy Machon<br><a href=mailto:Andy@machona.freeserve.co.uk>Andy@machona.freeserve.co.uk</a><br><a href= > </a><br>

 

[dirtman]

The problem with analyzing slopes with zero cohesion in an effective stress analysis is that the critical failure will always be an arc with a very shallow, high radius shape. This is in effect the safety factor you will get from an infinite slope analyses simply using equations for that purpose as shown in Lambe's textbook. So, unless you simply want to examine the infinite slope stability, you will need to include a small c value.