Invalid parameter set for Hardening Soil with m = 1 and phi' < 20 degrees

Hello,

When using the Hardening Soil (or HS-ss) model I sometimes get an error when modelling soft or organic clays (see figure below) and Plaxis doesn't accept my parameter set.

After some trial and error I found out that it is the combination of the following parameters that causes most of the trouble: m = 1.0 in combination with a friction angle (phi') < 20 degrees.

I therefore have two questions regarding the modelling of a soft/organic clay with the hardening soil model:

Am I using realistic soil parameters for modelling a soft/organic clay?: From the Plaxis Material Models Manual I understand that for soft clays the power m should be taken equal to 1.0 (paragraph 6.1).  I am therefore wondering if it is very uncommon to use a friction angle of < 20 degrees for a soft/organic clay when modelling in Plaxis? (other modelparameters I use are for example: yunsat/ysat = 15,  E50 =4.000, Eoed = 2.000 , Eur = 24.000, c'ref = 0, K0-nc = 0,6744 (default))

If it is not uncommon to use m = 1 in combination with phi < 20 degrees: What would be the best way to model a soft/organic clay in a realistic way. The error message suggests that I should change the Eoed, E50, Eur or K0-nc parameters, but this does not make a lot sense to me because I have to increase or decrease the parameters a lot, which dramatically changes the stiffness behavior of the soil.

Thanks in advance for your help!

Parents
  • Dear J Hopman,

    Using m = 1 with a friction angle of 20 degrees is indeed fine for soft clays....the problem is in the ratio between E50/Eoed as the HS model doesn't like this ratio to be higer than approximately 2. Depending on the other parameters sometimes a little bit higher than 2 is still possible, sometimes a ratio of 2 is already too much.
    Typically these ratios of E50/Eoed occur for very soft clays and  that is why we in general say the the HS and HSsmall models are less suitable for very soft clay.
    Since very soft clays have most deformation from compression rather than shear, a good alternative is to switch to the Soft Soil model - or, considering very soft clay often has creep behaviour, the  Soft Soil Creep model.
    So in short, we generally try to use the HS/HSsmall model, but if for very soft clays we can't enter the parameters we use Soft Soil (Creep) as plan B.

    With kind regards,

    Dennis Waterman

Reply
  • Dear J Hopman,

    Using m = 1 with a friction angle of 20 degrees is indeed fine for soft clays....the problem is in the ratio between E50/Eoed as the HS model doesn't like this ratio to be higer than approximately 2. Depending on the other parameters sometimes a little bit higher than 2 is still possible, sometimes a ratio of 2 is already too much.
    Typically these ratios of E50/Eoed occur for very soft clays and  that is why we in general say the the HS and HSsmall models are less suitable for very soft clay.
    Since very soft clays have most deformation from compression rather than shear, a good alternative is to switch to the Soft Soil model - or, considering very soft clay often has creep behaviour, the  Soft Soil Creep model.
    So in short, we generally try to use the HS/HSsmall model, but if for very soft clays we can't enter the parameters we use Soft Soil (Creep) as plan B.

    With kind regards,

    Dennis Waterman

Children
  • Dear D. Waterman,

    Thank you very much for your answer! (Sorry for my late response).

    Using the Soft Soil model is definitely a solution.

    However I do have the experience that also with a ratio E50/Eoed smaller than 2, Plaxis will still not accept my parameterset. But as soon as I change the m parameter to a lower value, for example 0,9, there is no problems at all. Using m = 1 in combination with the hardening soil model is therefore most of the times not succesfull for me (of course using the Soft Soil model is solution).

    Because of this I am actually wondering how often I should choose for a value of m = 1,0. Is it for example quite common to encounter a soil layer (in the Netherlands) that should be modelled with m= 1 or should m =1 only be used very (very) soft clays? I don't know if you have any thoughts (or expericence) about this?

    Kind regards,

    Jessy Hopman

  • Dear Jessy,

    The choice of m is not related to how soft the clay is but to the granulary content of the soil. Granular materials like sand tend to have values of about m=0.5 and non-granular materials like clay have values around m=0.9. And mixtures of soil and clay are somewhere in between. So choosing m=0.9 for clay is perfectly fine. Of course one should calculate m from an oedometer test rather than just assuming m=1.

    Traditional models assume that for clays there is a logarithmic stress-strain behaviour, and in the Hardening Soil model is simulated by m=1. That is in fact why a lot of people choose m=1.
    The difference between m=0.9 and m=1 is very small. Let's assume we have a saturated dutch clay with a specific weight of 15 kN/m3, than at 20 meters depth sig_1 = 100 kPa, so Eoed = Eoed,ref. At 10m depth the stiffness will be 0.5*Eoed,ref when m=1 is chosen and 0.535*Eoed when m=0.9 is chosen, so only 7% higher. 

    Note that the maximum ratio of E50/Eoed is not a fixed number, but it depends on several other parameters of which the strongest correlation is with K0nc. And then K0nc depends on the friction angle, hence of low friction angles (thus high K0nc) the maximum E50/Eoed relation may indeed be smaller than 2.  

    With kind regards,

    Dennis Waterman

  • Dear Dennis,

    Thanks again for your elaborate answer. This definitely helps me, it is a lot more clear to me now.

    For now that is all I need to know

    Kind regards,

    Jessy Hopman