Validating the Use of Material Point Method and SANISAND Model for

Relating the State Parameter with Cone Tip Resistance

Sara Moshfeghi, M.Sc., S.M.ASCE1, Mahdi Taiebat, Ph.D., P.Eng., M.ASCE2, and Arcesio
Lizcano, Ph.D., M.ASCE3
1
Department of Civil Engineering, University of British Columbia, Vancouver, BC, Canada; E-
mail: saramosh@mail.ubc.ca
2
Department of Civil Engineering, University of British Columbia, Vancouver, BC, Canada; E-
mail: mtaiebat@civil.ubc.ca
3
SRK Consulting Inc., Vancouver, BC, Canada; E-mail: alizcano@srk.com

ABSTRACT

Numerical simulations of Cone Penetration Test (CPT) can provide valuable insights into the
mechanical behavior and in-situ state of geomaterials. However, adequate simulation of CPT is
challenging due to the large deformations occurring during penetration and the need for
representative soil constitutive models. This study investigates the use of the Material Point
Method (MPM) and a version of the SANISAND family of models as a representative constitutive
model to simulate CPTs in dry sand. The numerical model is validated against experimental data
on CPT in a calibration chamber. The simulations are done on a range of soil overburden pressures
and void ratios. The results illustrate the sufficient adequacy of the numerical configuration in
capturing the cone tip resistance for the soil states considered. Additionally, the variation of void
ratio versus mean effective stress in selected material points is presented to explore details of
material state evolution during the cone penetration.

INTRODUCTION

The Cone Penetration Test (CPT) is a widely used in-situ test in geotechnical engineering to
characterize soils for a wide range of applications. However, the soil characterization approaches
based on CPT measurements have limitations due to the empirical nature of the correlations used.
Recent advancements in detailed numerical simulations of CPT have contributed to the
improvement of these correlations, leading to more accurate interpretation approaches (Ghasemi
et al., 2018; Jia et al., 2021; Zhang et al., 2021). Accurate simulation of CPT requires two crucial
components: representative constitutive models that can capture the nonlinear response of soil and
a robust computational method capable of handling large deformation problems.
The Material Point Method (MPM) is a continuum-based numerical framework, originally
proposed by Sulsky et al. (1994, 1995) that combines the benefits of point-based methods and
mesh-based procedures, making it suitable for large deformation simulations (Fern et al., 2019). It
is well-suited for large deformation problems, such as CPT. Some recent studies have employed
the MPM framework to simulate cone penetration, yielding promising results (e.g., Bisht et al.,
2021; Ceccato et al., 2016; Martinelli and Galavi, 2021; Yost et al., 2022). These examples
demonstrate the potential of MPM to simulate cone penetration accurately.

–1–
 In the study of numerical modeling of CPT, Kouretzis et al. (2014a, b) highlighted the
importance of the soil constitutive model employed in the simulations. They used an Arbitrary
Lagrangian-Eulerian (ALE) re-meshing technique and a critical state sand model that followed the
hardening law of the Modified Cam-Clay model. The constitutive model is expected to capture the
dependence of cone resistance on the effective stress, relative density, and compressibility
characteristics of the soil deposit. Furthermore, due to the large shear strains experienced by the
soil during cone penetration, it is crucial to use advanced soil constitutive models that respect the
critical state soil mechanics framework, for a realistic stress-strain response. Bisht et al. (2021), in
their study on numerical modeling of CPT based on MPM using a bounding surface constitutive
model for clays, also pointed out the importance of using an adequate constitutive model for such
simulations, highlighting the key relevant aspects of clay behavior: (a) the response during
consolidation, including loading-unloading curves in the void ratio-mean effective stress space,
(b) the transition from peak to residual state, and (c) the strain-rate dependency. Although these
factors hold true for predominantly clay-like soils, one needs to account for additional key features
when dealing with sand- and silt-type soils, as well as tailings, where the response can be more
intricate. In such cases, the material state, which includes stress level, void ratio, and various fabric
measures, is crucial in determining the contractancy/dilatancy and hardening/softening response
when the material is subjected to shear loading during the CPT. Bounding surface models have
proven reasonable success in handling stress-strain nonlinearity. In the modeling of CPT, other
crucial factors, such as fabric and anisotropic hardening, as well as the use of an adequate flow
rule to account for material shear-volume coupling, are expected to play a key role.
This paper aims to validate the simulation of CPT in dry sand using the well-known stress-
ratio controlled, critical state compatible, bounding surface plasticity model proposed by Dafalias
and Manzari (2004), referred to as DM04, within the MPM-based platform Anura3D. The
simulations are performed for homogeneous profiles of a given sand and at various initial void
ratios and overburden pressures. The results are then compared to those from chamber tests on the
same sand and at similar ranges of material state parameters.

EXPERIMENTAL TESTS

The experimental CPT data for validating the CPT simulation were from the study conducted by
Fioravante et al. (1991). The experimental data are obtained by calibration chamber tests
performed in Toyoura sand. The study involved sets of CPTs performed on Toyoura sand within
a calibration chamber. In total, 28 cone penetration tests were carried out using three different cone
diameters: 35.7 mm, 20 mm, and 10 mm. The calibration chamber had a diameter of 1.2 m and a
height of 1.5. The boundary conditions of the chamber were set as zero average strain and constant
stress in both vertical and horizontal directions. The specimens consisted of a range of void ratios
from 0.654 to 0.823, and were subjected to an overburden stress of 107 to 144 kPa, with the
majority of tests having an overburden stress of less than 122 kPa. The cone penetration velocity
was 2 cm/s, resulting in cone tip resistances between 4.3 to 40.4 MPa.

CONSTITUTIVE MODEL

The constitutive model, which formed the basis of what was later named SANISAND class of
models (Taiebat and Dafalias, 2008), accounts for material anisotropy by incorporating the Lode

–2–
angle dependency of the reference surfaces, including critical stress ratio, dilatancy stress ratio,
and bounding stress ratio, and also by kinematic hardening of the yield surface.
The plastic module and dilatancy of the model are informed by the concept of state
parameter 𝜓 = 𝑒 − 𝑒𝑐 , where 𝑒 and 𝑒𝑐 refer to the current void ratio and critical state void ratio
corresponding to the current mean effective stress 𝑝, respectively. This feature, in principle, is
designed to allow for modeling the hardening/softening and contractancy/dilatancy aspects of
response for different states of the material. Depending on the sign of 𝜓 the model adjusts the
reference bounding and dilatancy stress ratios, which in turn affects the corresponding distances
to the stress ratio. These distances, together with the model constants ℎ0 and 𝐴0 , control the plastic
modulus and dilatancy, respectively. At 𝜓 = 0, the bounding and dilatancy surfaces collapse onto
the fixed critical state surface, i.e., the critical state ratio in 𝑝 − 𝑞 space. Additionally, the model
incorporates a fabric dilatancy tensor that accounts for the fabric changes during plastic dilation
and its effect on the subsequent contraction upon reverse loading, a relevant feature for dealing
with cyclic shearing. The model parameters can be calibrated in a fairly straightforward manner
using standard types of laboratory tests. For example, drained and undrained triaxial compression
and extension tests at different values of initial void ratio and confining pressure allow for
calibrating different features of hardening/softening and dilatancy/contractancy of the model, with
a step-by-step calibration process presented in Taiebat et al. (2010). Table 1 presents the model
parameters for Toyoura sand, adopted for the presented study from Dafalias and Manzari (2004).

Table 1. Constitutive model constants for Toyoura sand

Description Symbol Value Description Symbol Value
Elasticity 𝐺0 125 Plastic modulus ℎ0 7.05
𝜈 0.05 𝑐ℎ 0.968
Critical state 𝑀𝑐 1.25 𝑛𝑏 1.1
𝑀𝑒 0.89 Dilatancy 𝐴0 0.704
𝜆𝑐 0.019 𝑛𝑑 3.5
𝑒𝑐_𝑟𝑒𝑓 0.934 Fabric dilatancy 𝑧𝑚𝑎𝑥 4
𝜉 0.7 𝑐𝑧 600

CPT MODEL DESCRIPTION

A 2D-axisymmetric numerical model was developed using the ANURA3D platform

(Anura3D, 2022), following the details presented in Moshfeghi et al. (2023). Simulating the exact
dimension of the calibration chamber would have made the simulations computationally
expensive. Instead, the model width was determined to be 0.32 m, around 20 times the cone radius
(20 𝑟c ) based on preliminary sensitivity studies. The height of the soil domain was set to 1.0 m,
and the cone penetration depth was limited to 0.4 m in order to avoid any effect from the bottom
boundary on the results. The model boundary conditions included vertical and horizontal
displacement constraints at both the top and the bottom boundaries. For the left and right
boundaries, horizontal displacements in the direction normal to the boundary were fixed.
Simulations were performed at two different overburden stress of 107 and 122 kPa. These values
were selected as they were the minimum and maximum overburden stresses in the experiment.
Also, the simulations were carried out at three different state parameters -0.25, -0.20, and -0.15.

–3–
The selection of these states was meant to cover a range of available data in the experiment.
Therefore, overall six simulations were performed considering two different overburden stresses
and three different initial state parameters. The overburden stresses were applied as external
traction at the top of the soil, and stresses were initialized using the K0 procedure with the K0 value
of 0.5. In accordance with standard cone specifications, the cone had a radius of 3.17 cm with an
apex angle of 60 degrees and was embedded in the soil to a depth of 3.75 cm. The simulated cone
diameter was close to the largest of the three diameters used in the chamber test. Figure 1 illustrates
the model’s configuration and additional details, including the dimensions, reference coordinate
system at the cone tip, boundary conditions, compressing and moving mesh, discretization, and
the number of material points per element (MP/EL) in different parts of the soil domain.
To prevent numerical instabilities, the tip of the cone was slightly rounded at its connection to the
shaft. The cone was modeled as a rigid body; hence no stress-strain relationship was considered
for the cone material points, and the cone only moved downward at the prescribed velocity. The
contact between the cone and soil was governed by the algorithm proposed by Bardenhagen et al.
(2001), with an interface friction angle of 31.5 degrees. To reduce stress oscillations, the
homogenous local damping of 10% was employed. Furthermore, to prevent the locking of
elements, the strain smoothing technique proposed by Al-Kafaji (2013) was used. The DM04
model numerical implementation follows the Runge-Kutta-Dormand-Price integration scheme. To

Figure 1. Configuration of the numerical model

–4–
ensure accurate numerical integration of the DM04 model, a time increment of 1E-4 sec was
chosen for all simulations. The moving mesh technique was used for spatial discretization. This
technique maintains a fine mesh around the cone tip throughout the simulation and ensures a well-
defined geometry in the contact zone between the cone and soil, making computations more
efficient. A finer mesh was used at closer distances to the centerline of the cone, while the mesh
became coarser further away from the cone to reduce computational cost. The boundary between
moving and compressing mesh is 15 cm below the top of the soil (12 cm below the cone tip). The
number of material points per element is a critical factor in ensuring that no empty elements are
present during simulation, which could lead to errors in the final output. Additionally, a higher
resolution is preferred in the vicinity of the cone. Therefore, the number of material points per
element in the soil close to the cone tip was set to 25 MP/El. To reduce the computational cost, the
number of material points per element was reduced to 12 and 6 in the zones progressively further
from the centerline of the cone.

RESULTS AND DISCUSSION

Figure 2 presents the cone tip resistance values obtained from the simulations for different
combinations of initial vertical stresses and initial void ratios. The values were determined based
on the vertical reaction forces recorded at the cone tip during the penetration. It should be noted
that the reported penetration depth 𝑧 in this and subsequent figures is measured with respect to the
initial position of the cone tip at the beginning of the penetration, based on the coordinate system
shown in Figure 1. This excludes the initial embedment of 3.75 cm below the soil surface.

Figure 2. Results of cone tip resistance profiles obtained from simulations

–5–
 Figure 3. Variations of the material state in e-p space at the three control material points
for simulation with overburden stress of 122 kPa and initial state parameter of -0.20

Three specific material points CMP 1, CMP 2, and CMP 3 were selected as control points
(CMPs) to analyze the changes in mean effective stress and void ratio during cone penetration.
The initial positions of the CMPs were located at a depth of 𝑧 = 18 𝑟c , with radial distances of 𝑟 =
1.0 𝑟c , 2.0 𝑟c , and 3.0 𝑟c from the initial position of the cone tip, which are referred to as CMP 1,
CMP 2, and CMP 3, respectively. The variations of void ratio versus mean effective stress at each
CMP during penetration are presented in Figure 3 for the simulation with the overburden stress of
122 kPa and initial state parameter of -0.20. The figure shows the same initial states for the three
CMPs, and they follow different paths as the penetration proceeds. According to this figure,
CMP1, the closest control material point to the cone, experiences an increase in mean effective
stress and dilation as the cone approaches the elevation of the CMP. This continues till the cone
tip almost reaches the elevation of the CMP where the CMP crosses the Critical State Line (CSL).
Once the cone tip reaches the depth of the CMP 1, following the CSL crossing, there is a notable
increase in mean effective stress and a significant decrease in the void ratio with further cone
penetration. When the cone sufficiently passes the depth of the CMP 1, the material at that location
experiences notable dilation and a drop in the mean effective stress.
The path of CMP 2 in Figure 3 represents the evolution of the material state further from
the cone sleeve when the cone tip reaches the depth of the control points. CMP 2 initially follows
a similar path to that of CMP 1, showing an increase in dilation and mean effective stress from the
beginning of the penetration. However, the key difference is that the CMP 2 path gets closer, but
does not reach the CSL. The mean effective stress at CMP 2 reaches its maximum when the cone
tip reaches the depth of the control points. As the cone surpasses the depth of the control points
and continues its penetration, the mean effective stress decreases at the CMP 2 location, while the
void ratio continues to increase until the cone tip reaches a depth of 0.4 m. CMP 3 is the furthest
material point, and as can be seen, it does not reach the critical state line and experiences a drop in
the mean effective stress and dilation as the cone passes the elevation of the CMPs.
Finally, Figure 4 illustrates the variation of cone tip resistance versus the initial void ratio
for the experiment and simulations. The values of the cone tip resistance shown in this figure are
obtained based on the calculated cone tip resistance values corresponding to the penetration depth

–6–
 Figure 4. Comparison of experimental and simulated cone tip resistance versus initial
void ratio

of 0.4 m in the profiles illustrated in Figure 2. It can be observed from the experimental data that
as the value of the void ratio increases, the measured cone tip resistance decreases. The six
simulations performed in this study sufficiently cover the range of void ratios considered in the
experiment. Comparing the experimental data and the simulation results shows that the simulated
values follow the same trend as the experimental data. The simulations for denser cases match well
with the experimental data. However, for the looser cases, the simulations tend to overestimate the
cone tip resistance slightly. Overall, it can be observed that although the cone diameter is different
from the ones used in the experiment, the results obtained from the simulations are in good
agreement with the measured values in the experiment, especially for denser cases.

CONCLUSIONS

The primary focus of the study was to validate the simulation of cone penetration tests by
comparing the results with physical modeling data obtained from a calibration chamber. The
simulations were conducted using the base version of the SANISAND family of models (DM04)
in the Anura3D platform. The model parameters and numerical settings were described in detail
to ensure the transparency and reproducibility of the simulations. Following the description of the
MPM model, the profile of cone tip resistance results was presented for all simulated cases,
providing an overview of the simulated behavior under different initial states regarding void ratio
and overburden pressure. Subsequently, the evolution of the material state at certain representative
control points was analyzed in detail for a selected simulation, allowing for a critical state-based
understanding of the behavior during the cone penetration process. Finally, the simulated cone tip
resistance values for all simulations were compared to the corresponding values obtained from
cone penetration tests performed in the calibration chamber. The successful validation was
demonstrated by the close agreement between the simulated cone tip resistance and the physical
modeling data, indicating the accuracy of the simulation in replicating the observed behavior
during cone penetration tests.

–7–
ACKNOWLEDGMENTS

Support for this study was provided by Mitacs, SRK Consulting Inc., and Industrias Peñoles.

REFERENCES

Al Kafaji, I.K., 2013. Formulation of a dynamic material point method (MPM) for geomechanical
problems. Ph.D. thesis, Dept. of Civil and Environmental Engineering, University of
Stuttgart.
Anura3D. 2022. Anura3D MPM research community. Online-External Workshop & Official
Release, accessed on April 22, 2022, http://www.anura3d.com/.
Bardenhagen, S.G., Guilkey, J.E., Roessig, K.M., Brackbill, J.U., Witzel, W.M. and Foster, J.C.,
2001. Improved contact algorithm for the material point method and application to stress
propagation in granular material. Comput. Modell. Engineering Science, 2(4), 509-522.
Bisht, V., Salgado, R. and Prezzi, M., 2021. Material point method for cone penetration in
clays. Journal of Geotech and Geoenv Engineering, 147(12), 04021158.
Ceccato, F., Beuth, L., Vermeer, P.A. and Simonini, P., 2016. Two-phase material point method
applied to the study of cone penetration. Computers and Geotechnics, 80, 440-452.
Dafalias, Y.F., Manzari, M.T., 2004. Simple plasticity sand model accounting for fabric change
effects. Journal of Engineering Mechanics, 130(6), 622-634.
Fern, J., Rohe, A., Soga, K. and Alonso, E., 2019. The material point method for geotechnical
engineering: a practical guide. CRC Press.
Fioravante, V., Jamiolkowski, M., Tanizawa, F. and Tatsuoka, F., 1991, June. Results of CPTs in
Toyoura quartz sand. In Proceedings of the 1st international symposium on calibration
chamber testing (ISOCCT-1), Potsdam, New York, Elsevier (pp. 135-146).
Ghasemi, P., Calvello, M., Martinelli, M., Galavi, V. and Cuomo, S., 2018. MPM simulation of
CPT and model calibration by inverse analysis. In Cone Penetration Testing, 295-301,
CRC Press.
Jia, R., Zhao, D., Lei, H. and Li, Y., 2021. Numerical study of the effect of clay structure on
piezocone penetration test results. International Journal of Geosynthetics and Ground
Engineering, 7, 1-14.
Kouretzis, G.P., Sheng, D. and Wang, D., 2014a. Numerical simulation of cone penetration testing
using a new critical state constitutive model for sand. Computers and geotechnics, 56, 50-
60.
Kouretzis, G.P., Sheng, D.C. and Wang, D., 2014b. Numerical simulation of CPT cone penetration
in sand. In Applied Mechanics and Materials, 553, 416-421, Trans Tech Publications Ltd.
Martinelli, M. and Galavi, V., 2021. Investigation of the material point method in the simulation
of cone penetration tests in dry sand. Computers and Geotechnics, 130, 103923.
Moshfeghi, M., Taiebat, M. and Lizcano, A. (2023). The role of soil constitutive model in
simulation of cone pen- etration test. Proceedings 10th NUMGE 2023: 10th European
Conference on Numerical Methods in Geotechnical Engineering, Zdravkovic L., Konte,
S., Taborda, D.M.G., Tsiampousi, A., eds., London, UK, Paper ID: 252, 6 pages.
Shuttle, D. and Jefferies, M., 1998. Dimensionless and unbiased CPT interpretation in
sand. International Journal for Numerical and Analytical Methods in Geomechanics 22(5),
351-391.

–8–
Sulsky, D., Chen, Z., Schreyer, H.L., 1994. A particle method for history-dependent
materials. Computer Methods in Applied Mechanics and Engineering, 118(1-2), 179-196.
Sulsky, D., Zhou, S.J., Schreyer, H.L., 1995. Application of a particle-in-cell method to solid
mechanics. Computer Physics Communications, 87(1-2), 236-252.
Taiebat, M., Dafalias, Y.F., 2008. SANISAND: simple anisotropic sand plasticity model.
International Journal of Numerical and Analytical Methods in Geomechanics 32(8), 915-
48.
Taiebat, M., Jeremić, B., Dafalias, Y.F., Kaynia, A.M., Cheng, Z. 2010. Propagation of seismic
waves through liquified soils. Soil Dynamics and Earthquake Engineering, 30(4), 236-257.
Yost, K.M., Yerro, A., Green, R.A., Martin, E., Cooper, J., 2022. MPM modeling of cone
penetrometer testing for multiple thin-layer effects in complex soil stratigraphy. Journal of
Geotechnical and Geoenvironmental Engineering, 148(2), 04021189.
Zhang, W., Zou, J.Q., Zhang, X.W., Yuan, W.H. and Wu, W., 2021. Interpretation of cone
penetration test in clay with smoothed particle finite element method. Acta
Geotechnica, 16(8), 2593-2607.

–9–