Hy Professor David,
I am trying to verify the result of one of your paper by the title ‘Cyclic lateral response and failure mechanisms of semi-rigid pile in soft clay: centrifuge tests and numerical modelling’. The dimension of the model are same as described in the paper. Following step are performed in the numerical modelling procedure
1. OCR is initiated layer wise using props(22) with the same OCR pattern as described in your paper. (picture attached)
2. Initial acceleration of 40g is applied to the overall model to generate effective stress in the model.(picture attached)
3. Pore water pressure 0-180 Kpa is initiated both in the model as well as in SDVINI using state variable 8.(picture attached)
4. Based on the effective stress and OCR, using Bolton and Stewart 1994 relation i calculated under drained shear strength of the soil model.
5. Through this undrained shear strength the elastic modulus of the soil layer is calculated to be around 11600kpa i.e. 400 time of its undrained shear strength at mid height
6. Based on the relative pile-soil stiffness of 0.0067, the elastic modulus of pile calculated as 245 Mpa i.e. semi rigid pile modelling
In the first step of the model the geostatic stress are generated and the lateral displacement is applied to the top of the pile. Professor my question is that the displacement that i am obtaining is around 30mm for 18KN load and in your case its even less than 1mm(picture attached)…. Am i missing something in properties or modelling? Hope to hear from you soon. The detail of properties are attached. Thank you very much
Hi Ahmad, did you actually scale the applied force by centrifuge scaling laws? The graphs in the paper are shown in the prototype (not model) scale. See for example this paper for centrifuge scaling laws (https://www.researchgate.net/publication/344818529_Physical_modelling_of_piles_under_lateral_loading_in_unsaturated_soils/download)
Understood professor. Thank you very much . Few more question if you could please help me to understand
1. The effective stress in my case after apply 40g acceleration is almost twice i.e. approx. 100kpa in paper and 200 Kpa in my case. The specific weight of 1.65 kN/m2 is mentioned in paper. How come we can resolve this issue ? Does scaling laws applicable in this case as well ?
2. In the paper its written that pore water pressure of 0 kPa is applied to the top and 180 kPa at the bottom. There are two option of doing this, one is applying pore water pressure ass a function of height in software . Second is through initiating SDV8 as a function of height. i.e. SDV8=180kPa*coords(3). Which one should we adopt ?
3. How to define initial value of intergranular strain tensor in this case ?
Hi,
1. “Specific gravity” 16.5 kN/m3 is obviously an error, this is total unit weight. Depth in prototype scale is 18m, so effective vertical stress (16.5-10)*18 kPa=117 kPa
2. SDV8 is just for undrained analysis governed by umat through penalty approach. You should run consolidation analysis, with pore water pressures controlled by FEM software. Actually, SDV8 will remain zero all the time as umat will only receive information on effective stress.
3. If you simulate gravity acceleration, set them to 0 initially. If you start from stresses calculated using K0-procedure, intergranular strain should be initialised to -R for vertical normal component and 0 otherwise.
Regards David
Dear Professor, Thank you very much.it really helped. But i am still facing issue as the model is not showing stiffness degradation under cyclic load and normalized Pile head displacement is constant as the number of cycle increase . These are the few issues in which i would like your guidance further
1. In your book you have mentioned that OCR based initiation is usually preferred for softer clay. I have used OCR based initiation but i have read somewhere in comments that undrained shear strength is different for OCR based initiation and void ration based initiation. Would that be a Problem ? Can we show the value of undrained shear strength as state dependent variable?
2. In your online documentation it is written that if initial value of sensitivity is equal to 0 i.e. SDV14=0 than basic model is used. its mean basic hypoplastcity without intergranular stiffness. IS there a need to define initial value of sensitivity in y case ?
3. In the paper it is written that that ” This may imply that the soil simulated by the hypoplastic clay model in this study behaved more stiffly than it did in the centrifuge model test” I ma facing same issue as my lateral pile displacement is too small . Would to like to comment on this ?
4. As i the stresses in the model were initialized using K0 initialization procedure i.e. be defining the initial geostatic stress and then applying the external gravity . So i have used the following initialization technique for state dependent variables..
Subroutine SDVINI(STATEV,COORDS,NSTATV,NCRDS,NOEL,NPT,LAYER,KSPT,props,nprops)
Implicit Double Precision (A-H, O-Z)
Dimension statev(*), props(*)
c intialize state variables
statev(1)=0.d0
statev(2)=0.d0
statev(3)=0.0001
statev(4)=0.d0
statev(5)=0.0001
statev(6)=0.0001
statev(7)=props(22)
statev(8)=0.d0
statev(9)=0.d0
statev(10)=0.d0
statev(11)=0.d0
statev(12)=0.d0
statev(13)=0.d0
statev(14)=0.d0
statev(15)=0.d0
statev(16)=0.d0
Is that Okay ?
5. Need your guidance on this line ”Even though pile–soil tension (due to negative excess pore pressure) at the rear of the pile was ignored in the numerical analysis ”
6. In order to initialize the state dependent variables mainly the intergranular stiffness , in your SDVININ file there is loop that call SDV1 to SDV6 through Props 22. Can you Please explain whats the role of this loop in the model .
Here it is
if (statev(1) .eq. 0.0d0 .and.
* statev(2) .eq. 0.0d0 .and.
* statev(3) .eq. 0.0d0 .and.
* statev(4).eq. 0.0d0 .and.
* statev(5).eq. 0.0d0 .and.
* statev(6).eq. 0.0d0) then
do i=1,6
statev(i)=props(22+i) ! initialize e3_e3s
enddo
end if
if (statev(14) .lt. 0.001) then
statev(14)=props(29)
end if
Thank you very much professor for your time and effort.
Arsalan