Академический Документы
Профессиональный Документы
Культура Документы
AVISA: Anisotropic Visco ISA model and its performance at cyclic loading
--Manuscript Draft--
Full Title: AVISA: Anisotropic Visco ISA model and its performance at cyclic loading
Keywords: soft soils; clay; silt; cyclic loading; viscosity; anisotropy; rate-dependency; time-
dependency; excessive pore water pressure; partially flow
Abstract: In this work a constitutive model with some interesting capabilities, such as the
simulation of the strain rate and time dependency, the incorporation of small strain and
of fabric effects without restriction on the overconsolidation ratio is proposed. The small
strain effects of clays has been simulated through an extended ISA plasticity
formulation (first version of ISA plasticity has been proposed for sand by Fuentes &
Triantafyllidis). The rate dependency is captured by a third strain mechanism (in
addition to the total and the hypoplastic strain) that is dependent on the stress and the
void ratio. The loading surface has been well defined in order to provide a clean and
complete formulation for the overconsolidation ratio. The degree of nonlinearity
presented in the reference model [2] turned to be unsuitable as it does not provide
information about the preloading state of the material. For this purpose it has been
reformulated to account on the influence of OCR. Some investigations regarding the
inherent anisotropy has shown that the effective stress path in p - q space prior to the
appearance of the cyclic mobility is signicantly inclined towards the left upper corner of
the p - q plane for Kaolin Clay and Lower Rhine Clay and that the inherent anisotropy
of clays can be described through the transversal isotropy. Consequently the
hypoelastic stiffness has been transformed to a transversal hypoelastic one.
Furthermore, the last cycles at cyclic mobility show a prevented dilatancy at loading
and a contractant behaviour after an elastic regime is overcomed during unloading. At
stress reversal both clays first exhibited an increase of the mean pressure in contrast
to sand whereas the stress path on reversal starts with the highest contractancy. This
can present a reason why most clays do not reach liquefaction p = 0 kPa due to cyclic
loading. The performance of the model has been inspected by simulating the database
with about 50 cyclic undrained triaxial (CUT) tests on the low plasticity Kaolin Clay
published by Wichtmann & Triantafyllidis including variation of the deviatoric stress
amplitude, initial stress ratio, displacement rate, overconsolidation ratio and cutting
direction and 4 CUT tests performed on the high plasticity Lower Rhine Clay, whereby
the influence of the displacement rate as well as the deviatoric amplitude has been
studied. The simulations showed to be in a very good agreement with the experimental
observations.
Powered by Editorial Manager® and ProduXion Manager® from Aries Systems Corporation
Click here to download Manuscript AVISA_TT.pdf
1
2 AVISA: Anisotropic Visco ISA model and its performance at cyclic loading
3
4 Merita Tafilia,∗, Theodoros Triantafyllidisa,b
5
a Institute
6 of Soil Mechanics and Rock Mechanics (IBF), Karlsruhe Institute of Technology, Karlsruhe 76131, Germany
b Professor and Director of IBF
7
8
9
10
11
12 Abstract
13 In this work a constitutive model with some interesting capabilities, such as the simulation of the strain rate and time
14 dependency, the incorporation of small strain and of fabric effects without restriction on the overconsolidation ratio is
15
proposed. The small strain effects of clays has been simulated through an extended ISA plasticity formulation (first
16
version of ISA plasticity has been proposed for sand by Fuentes & Triantafyllidis). The rate dependency is captured
17
18 by a third strain mechanism (in addition to the total and the hypoplastic strain) that is dependent on the stress and
19 the void ratio. The loading surface has been well defined in order to provide a clean and complete formulation for the
20 overconsolidation ratio. The degree of nonlinearity presented in the reference model [2] turned to be unsuitable as it does
21 not provide information about the preloading state of the material. For this purpose it has been reformulated to account
22 on the influence of OCR. Some investigations regarding the inherent anisotropy has shown that the effective stress path
23 in p − q space prior to the appearance of the cyclic mobility is significantly inclined towards the left upper corner of the
24 p − q plane for Kaolin Clay and Lower Rhine Clay and that the inherent anisotropy of clays can be described through
25 the transversal isotropy. Consequently the hypoelastic stiffness has been transformed to a transversal hypoelastic one.
26 Furthermore, the last cycles at cyclic mobility show a prevented dilatancy at loading and a contractant behaviour after
27 an elastic regime is overcomed for unloading. At stress reversal both clays first exhibited an increase of the mean pressure
28 in contrast to sand whereas the stress path on reversal starts with the highest contractancy. This can present a reason
29 why most clays do not reach liquefaction p = 0 kPa due to cyclic loading. The performance of the model has been
30 inspected by simulating the database with about 50 cyclic undrained triaxial (CUT) tests on the low plasticity Kaolin
31 Clay published by Wichtmann & Triantafyllidis including variation of the deviatoric stress amplitude, initial stress ratio,
32 displacement rate, overconsolidation ratio and cutting direction and 4 CUT tests performed on the high plasticity Lower
33 Rhine Clay, whereby the influence of the displacement rate as well as the deviatoric amplitude has been studied. The
34
simulations showed to be in a very good agreement with the experimental observations.
35
36 Keywords: soft soils, clay, silt, cyclic loading, viscosity, anisotropy, rate-dependency, time-dependency, excessive pore
37 water pressure
38
39
40 1. Introduction On the other side, most natural clays show anisotropic
41 behaviour due to their mode of deposition and the elon-
42 Structures subjected to cyclic loading experience ir- gated shape of the particles [6, 7, 8]. This anisotropy,
43 reversible displacements resulting from the accumulated known as inherent anisotropy, along with the time depen-
44 strains within the underlying soil. A rapid cyclic load- 20 dent phenomena of soft soils have practical consequences
455 ing of a fully saturated soil leads under undrained condi- for i.e. the passive lateral thrust on piles, which has been
46 tions to a significant reduction of the effective stress. For the target of many investigations in the last decades. By
47 an accurate prediction of the settlement and the bearing
48 creeping slopes or construction of an embankment on such
capacity of an installed structure as for example a vibro- formations, passive time-dependent lateral pressure may
49
replacement with stone columns the cyclic behaviour of the 25 occur to the pile shaft. This may lead to deformations
50
10 underground consisting of clay as well as its time depen- of the pile foundation and of the superstructure. Be-
51
52 dent behaviour is of high importance. The majority of an- sides these effects, the behaviour of clays depends not
53 alytical settlement calculation methods pertain to primary only on the material state, but also on the strain ampli-
54 settlement only, thus the same value of the settlement im- tude. Researchers [9, 3, 10, 11, 12, 13, 14] have found that
55 provement factor tends to be applied to both primary and 30 clays behave elastically only under very small strain am-
56
15 creep settlements [4, 5]. plitudes |ε1 | < 0.01 %. Under medium strain amplitudes
57 0.01 % < |ε1 | < 1 %, the so-called small strain effects,
58 namely the stiffness increase due to reversal loading and
∗ Correspondingauthor
59 the reduction of the plastic strain rate take place. Finally,
Email address: merita.tafili@kit.edu (Merita Tafili)
60
61 Preprint submitted to Journal of LATEX Templates March 5, 2019
62
63
64
65
35
1 when the soil is sheared under very large deformations, it without changing the model from non-viscous to viscous
2 tends asymptotically to the critical state, which may re- clays while keeping the same set of material parameters.
3 sult to a failure. Thus, in order to simulate each of these
4 effects the usage of an appropriate model is recommended
2. ISA-plasticity for the small strain stiffness
5 e.g. [15, 16, 17, 18, 19, 20]. Yet, each of these effects is
6
40 important for the investigations of the soft soil behaviour 95 In order to acurately describe the behaviour of gran-
7 in urban excavations with shoring walls subjected to a low ular soils under cyclic loading, in particular for a bet-
8 number of cycles. This cyclic loading may be attrubited ter prediction of the small strain stiffness and the shear
9 to the excavation shoring and backfilling processes or in modulus degradation curve, Niemunis and Herle [22] pro-
10 case of offshore foundations with a large number of cyclic posed the concept of the intergranular strain (IS). In 2015
11
45 loading due to wave and wind actions or for the long term100 Fuentes and Triantafyllidis [1] reformulated the intergran-
12 design of stone columns as soil replacement subjected to a ular strain and named it the ISA-model, hereafter denoted
13 large number of cycles carrying traffic loading. In this pa- as ISA-plasticity, introducing an elastic locus in the inter-
14 per a single model capturing all these effects is proposed.
15 granular strain space. Tab. 2 provides a summary of the
The paper examines the anisotropic material response of models developed unter the ISA-plasticity framework. In
16
50 clays resulting in transverse isotropy. Only one additional105 the following lines the ISA-plasticity formulation for sand
17
18 parameter is required to describe this type of inherent elas- [1, 23, 2, 24, 25] is summarized and its extensions to cap-
19 tic anisotropy [21]. Among all strain amplitudes the strain ture the specific behaviour of soft soils are described.
20 rate dependency of clays is experimentally investigated. The main feature of the ISA-plasticity formulation for sand
21 These effects are incorporated into the proposed model by is the incorporation of the elastic locus as a strain-type
22
55 an extension and modification of the (hypo)elastic stiff-110 yield surface. It describes a hypersphere with diameter R
23 ness tensor and the consideration of an additional strain and depends on the intergranular strain tensor h and the
24 mechanism, whose intensity depends on the material vis- kinematic hardening tensor c:
25 cosity as depicted in [2]. The model is proposed under
26 the platform of the ISA-plasticity, which has shown good FH = kh − ck − R/2 = 0 (1)
27
60 simulation results for cyclic loading and captures well the The evolution of the intergranular strain tensor is described
28 small strain effects for sands. These advantages of the in accordance with elastoplastic formulations introducing
29 ISA-plasticity has been extended and adjusted to the me-115 the consistency parameter λ̇H and the flow direction N:
30 chanical behaviour of clays.
31 The structure of this article is as follows: At the begin- ḣ = ε̇ − λ̇H N, (2)
32
65 ning, the formulation of the proposed model is explained.
33 whereby λ̇H fulfills the consistency condition ḞH = 0 and
Then, some details about the numerical implementation
34 is derived to:
and the integration scheme are given. The performance of
35 hN : ε̇i hN : ε̇i
the model is finaly illustrated in the case of undrained tri- λ̇H = = . (3)
36
axial cyclic simulations of two different clays: the medium N : N + N : c̄ 1 + HH
37
70
38 plasticity Kaolin Clay and the high plasticity Lower Rhine For the plastic condition FH = 0 an associated flow rule
39 Clay. The experiments for the Kaolin Clay are borrowed N = ∂FH /∂h and a kinematic hardening c of the yield sur-
40 from [3], whereas the experiments for the Lower Rhine120 face are introduced as a transition state until the bounding
41 Clay are performed by the authors and Wichtmann as will surface (BS) is reached:
42 be reported in a subsequent paper. The simulations are
43
75 carefully analyzed regarding the strain rate effects, dif- FHb = khk − d · R = 0. (4)
44 ferent overconsolidation ratios, various initial stress ratios
The BS has been formerly proposed in [1] according to
45 and mean preassures and different sample cutting direc-
the bounding surface plasticity characteristics [26] repre-
46 tions. The samples where subjected to various deviatoric
senting it similarly to the yield surface as a hypersphere
47 amplitudes.
48 125 in the IS-space, but with twice the diameter of the yield
80 The proposed model is able to describe the material be-
49 locus FHb = khk − R = 0. Following Eq. (4) we propose
haviour of viscous and non viscous clays under cyclic as
50 a changed size of the bounding surface through the new
well as monotonic loading, capturing also the present in-
51 parameter d. Hence, for sand this parameter can be set
herent anisotropy of clays. The model covers a wide range
52 to d = 1 yielding the formulation of [1]. For clays, the ex-
of strain amplitudes without restriction on the overcon-
53 130 perimental experience gained in this work has shown that
85 solidation ratio (OCR). Normal consolidated as well as
54 d tends to d = 2. The indroduction of a new size of the
slightly and heavily over consolidated clays behaviour can
55 bounding surface has required the adjustment of the fol-
be simulated. Simulations of tests with different initial
56 lowing equations compared to [1].
57 stress ratios show the good performance of the model in
The kinematic hardening tensor represents the center of
58 capturing the induced anisotropy of Kaolin Clay. From
135 the yield surface and reads:
59
90 the point of constitutive modeling, this is very interesting
60 because the proposed model is able to simulate each effect ċ = λ̇H c̄, (5)
61 2
62
63
64
65
1 In order to fulfill the bounding constraint khk ≤ R the [28] χ = χ0 + εa (χmax − χ0 )). Laboratory tests dealing
2 hardening mechanism should evolve towards the image165 with cyclic behaviour of clays with different number of
3 tensor at the BS (projection in direction of N) denoted cycles [11] has shown that the plastic accumulation rate
4 by cb . Thereby, the hardening function is proposed to reduces for increasing number of consecutive cycles before
5
140 read: reaching the critical state. This effect can be described
6 → through a new state variable εa [28], which distinguishes
7 c̄ = β (cb − c)/(d · R), cb = R (d − 1/2) ε̇ (6)170 whether the soil is performing a few or severeal consecutive
8 cycles. Its evolution rate is controlled through a material
whereby β = f (β0 , σ) is here proposed as a variable that
9 parameter Ca and reads:
controls the hardening rate as will be defined in the sub-
10
11 sequent section. For isotropic conditions it takes the value ε̇a = Ca /R (1 − yh − εa ) kε̇k. (10)
12 of the material parameter β(q = 0) = β0 .
13 Thus, for a few cycles εa ≈ 0 ⇒ χ = χ0 and the IS expo-
14 nent proposed by [1] is recovered. εa → 1 is reached when
15 175 several consecutive cycles have been experienced and the
16 exponent then increases to χ → χmax . χ0 and χmax are
17 material parameters that can be calibrated for a few and
18 several number of cycles, respectively.
19 The increase of the elastic stiffness at unloading is achieved
20 180 through the scalar function m = mR +(1−mR ) yh , whereby
21 mR is a material parameter introduced first in [22, 9].
22 If the stress loops are performed at the critical state, even-
23 though the strain rate may be small |∆ε| < R the material
24 yield surface
response is fully mobilized. To capture this effect we pro-
25 bounding surface 185 pose a relation for the scalar function β, which controls
26 the hardening rate of the IS (compare Eq. (6)):
27
28 3
Figure 1: Yield surface and bounding surface of the extended ISA β = β0 + (2 − β0 ) (krk/krc k) , (11)
29 plasticity for clays. Example when the intergranular strain is fully
30 mobilized h = hb and c = cb with the stress ratio r and its image at the critical state
31 rc .
32 The formulation of the general model might at first sight
33
145 2.1. Mechanical model 190 appear high, but under some specific assumptions it re-
34 According to the ISA-plasticity, for material states in- duces to some model families that possibly are simpler to
35 side the IS yield surface k∆εk < R the response of the understand, as outlined in Tab. 1. For example if the yield
36 model is visco-elastic thus λ̇H = 0 and the intergranular surface radius goes towards R → 0 than the ISA-plasticity
37 strain evolves equaly to the applied strain. The (visco- is disregarded and we obtain the constitutive model for-
38 elastic) constitutive equation for these states reads:
150 195 mulation under mobilized states.
39
vis
40 σ̇ = m E : ε̇ − ε̇ . (7)
41 3. Constitutive model under mobilized states
42 Once the yield surface is reached, the kinematic hardening
43 of the centre of FH activates c̄ with the IS consistency The fully mobilized state is reached when the inter-
44 parameter λ̇H , which defines also the plastic evolution of granular strain h lies at the IS bounding surface FHb =
45 the IS state variable h. The area between the yield surface 0 ⇒ yh = 1 ⇒ m = 1. Hence the constitutive equation for
46
155 and the bounding surface is termed as a trasition area200 clays at fully mobilized states reads:
47 between the visco-elastic Eq. (7) and the fully mobilized
σ̇ = E : ε̇ − ε̇hp − ε̇vis .
48 state yh = 1: (12)
49
σ̇ = m E : ε̇ − yh ε̇hp − ε̇vis .
50 (8) This relation is visco-hypoplastic as described in [21]. The
51 evolution equation of the hypoplastic strain rate ε̇hp reads:
The bounding condition h = hb i.e. the fully mobilized
52 ε̇hp = Y mkε̇k (13)
state should be asymptotically reached ensured with the
53
160 scalar function 0 ≤ yh ≤ 1:
54 and for the viscous strain rate ε̇vis :
55 → khb − hk
yh = ρχ hN : ε̇ i,
1/Iv
ρ=1− , (9)
56 d·R 1
ε̇vis = Iv λ m (14)
57 OCR
58 whereby hb = d/2·R N is the image tensor of the IS at the
59 bounding surface. The exponent χ accounts indirectly for with the compression index λ and the viscosity index Iv .
60 the number of constitutive cycles the soil has experienced205 The same direction is used for both plastic strains [24, 2,
61 3
62
63
64
65
1 Table 1: Simplification of the general constitutive model to some particular forms
2 Name Assumptions Simplified equation Remarks
3
4
5 Elastic R → ∞, Iv = 0 σ̇ = mR E : ε̇ time-independent, no accumulation
6 Visco-elastic R→∞ σ̇ = mR E : (ε̇ − ε̇vis ) time-dependent, no accumulation
7 Hypoplastic R → 0, Iv = 0 σ̇ = E : (ε̇ − ε̇p ) time-independent, monotonic loading
8
Visco-Hypoplastic R→0 σ̇ = E : (ε̇ − ε̇p − ε̇vis ) time-dependent, monotonic loading
9
10 ISA-plasticity none σ̇ = mE : (ε̇ − yh ε̇p − ε̇vis ) time-dependent, monotonic + cyclic loading
11
12
13 Table 2: Summary of the models developed unter the ISA-plasticity framework
14 Name Geomaterial Description Literature
15
16 ISA - Sand (2014) Sand Uses projections in the stress space (similar to [27]) [1]
17
18 ISA - HP Wolffersdorff (2015) Sand Without projections in the stress space [28]
19 ISA - HP Clay (2015) Clay Rate independent [23]
20 ISA - visco HP Clay (2016) Clay Inclusion of rate and time dependency [2]
21
22 ISA - aniso visco HP Clay (2017) Clay inherent anisotropy introduced this work
23
24
25 25, 21]: δil δjk ) The barotropic bulk modulus K of clays linearly in-
26 → creses with increasing mean stress p as shown in [2, 21]:
27 1 r
m = − (krc k − krk) + . (15) p (1 + e)
28 2 krc k K= (17)
29 λ (1 − Y0,max )
30 The equations governing the reference model [2] are sum-
marized in Appendix A. In this section the differences whereby Y0,max = (λ − κ)/(λ + κ) is the degree of non-
31
between the reference model and the AVISA model are235 linearity deducted for normal consolidated states upon un-
32
in particular presented. The new approach for the de- loading [2, 21]. The shear modulus G can be deduced from
33
210
34 scription of fabric effects, as well as the influence of the K and the Poisson ratio ν to read:
35 overconsolidation ratio (OCR), formulated with a param- 3 (1 − 2ν)
eter less than the required number of the reference model G=K . (18)
36 2 (1 + ν)
37 parameters, and the new relation for the degree of nonlin-
38
215 earity Y are described in detail. Most natural clays show anisotropic behaviour because
39 Furthermore it should be noted that experimental studies of their mode of deposition and the elongated shape of
40 [29, 30, 31] has shown that the viscous effects of (highly)240 the particles [21]. This anisotropy, known as inherent
41 overconsolidated clays vanish exponentially. For cyclic anisotropy or fabric, resulting from the deposition pro-
42 loading of a normalconsolidated sample this means that cess tends to induce a horizontal bedding plane in the soil
43
220 the viscous effects decrease exponentially with each cycle. layer and can be regarded as a transverse isotropy [6, 21].
44 In the constitutive Eq. (7) of the AVISA model the vis- For undrained monotonic loading the transverse isotropy
45
cous strain is active also for elastic conditions. Yet, Eq.245 influences the slope of the effective stress path. In case of
46 undrained cyclic loading the described elastic anisotropy
(14) suggests that its magnitude decreases with increasing
47 besides the effective stress slope affects the direction of
48 OCR exponentially. The exponent 1/Iv takes high values
225 for 0 ≤ Iv ≤ 0.5 and ensures the vanishing viscous strain the strain accumulation, whether it occurs in extension
49 or compression. Thus, a performance based design of a
50 for elastic conditions e.g. overconsolidated clays.
250 geotechnical structure is not possible without describing
51 the fabric well.
52 3.1. Inclusion of fabric effects for clays
Furthermore, experiments show that after a stress rever-
53 A hypoelastic stiffness has been introduced in the AVISA sal first an increase of the mean pressure occurs in most
54 model: clays in contrast to sand where the reversal starts with the
55
→→ K 255 highest contractancy. This can also present a reason why
56 Eiso = 3K 1 1 +2G Idev − √ (1 r + r 1) (16)
57 3Mc most clays do not reach liquefaction p = 0 kPa. Thus, also
58 →→
in order to capture the dilatancy/contractancy behaviour
59
230 with the fourth rank deviatoric unit tensor Idev = I− 1 1 of soils well, first the accurate description of elasticity and
60 and the unit tensor for symmetric tensors Iijkl = 1/2(δik δjl + of the inherent anisotropy is required.
61 4
62
63
64
65
260
1 For a complete transverse isotropy five material param- 3.3. Degree of nonlinearity Y
2 eters are required. However, according to the pioneer The degree of nonlinearity is proposed according to
3 work of Graham & Houlsby [6] for the special case of some hypoplastic requirements:
4 anisotropic clays a simplified transversal isotropic elastic-
5 ity is deduced, whereby only three parameters are needed: • isotropic states (q = 0): Y takes the minimum value
6
265 the Young’s modulus E = Ev , the Poisson ratio ν = νh 305 Y = Y0 , whereby the influence of OCR is also con-
7 and the anisotropic factor α obeying the relation: sidered: Y0 = Y0,max (p/pei )2 [2].
8
9 Gh /Gv = (Eh /Ev )1/2 = νh /νvh = α (19) • states lying at the loading surface Fb = 0 or at the
10 critical state surface Fc = 0: experimental observa-
11 The index h stands for the horizontal direction and v vor tions has shown that at these states the hypoplastic
12 vertical direction. Two of these material constants are310 degree of nonlinearity takes the value Y = 1.
13 already present in the proposed model with E = f (κ)
270
14 and ν. Thus, only one additional material parameter α • the degree of nonlinearity should provide information
15 needs to be introduced into the model. Now the question about the state of the material: the loading surface
16 arises how can a stiffness of a model be transformed to a provides the boundary where Y = 1 and inside the
17 transversal one? After some mathematical manipulations loading surface Y decreases towards Y = Y0 which
18 it becomes obvious that each hypoelastic stiffness can be315 is reached at the isotropic axis. This requirement is
19
275 transformed to a transversal one by scaling it according to firstly termed in this work.
20 the following schema [32]:
21 A suitable relation to fulfill these requirements is the fol-
22 Eabcd = Qabij : Eijkl : Qklcd (20) lowing equation:
23
OCR − 1
24 Qabcd = µac µbd (21) Y = Y0 + (1 − Y0 ) exp − . (24)
krk
25
26 √ √ For isotropic states ||r|| → 0 the relation (24) tends to:
27 µ= α1 + (1 − α)ms ⊗ ms ,
28 lim Y = Y0 (25)
ms is the unit vector along the sedimentation axis e.g. if r→0
29
280 the sedimentation axis is horizontal then ms = {1, 0, 0}.
30 320 thus, the first requirement is fulfilled. States lying at the
31 Hence, only 1 additional parameter termed as anisotropic
loading surface or at the critical surface are normalconsol-
coefficient α compared to the reference model [2] is re-
32 idated OCR = 1, hence Y → 1. The incorporation of the
33 quired.
loading surface as a boundary for the degree of nonlinearity
34 is obvious from Eq. (24). Both the degree of nonlinear-
35 3.2. Overconsolidation ratio OCR
325 ity proposed in Eq. (24) and the reference model’s degree
36
285 Whereas in [2] an interpolation function is used for of nonlinearity Yrefmod [2] are illustrated in Fig. 2. De-
37 OCR introducing an additional material parameter, a more spite the OCR-independence of the Yrefmod , it possesses
38 precise and comprehensive OCR formulation is proposed also numerical issues approaching the isotropic axis as de-
39 herein. Following Hvorslev [33] we introduce the equiva- picted in Fig. 2 a) with dashed green lines. The orange
40 lent isotropic pressure pei = exp((ei0 − e)/λ). After in-330 limit Y = Y0 holds for both definitions equally.
41
290 corporating a loading surface for the stress Fb = 0 where
42
OCR= 1 holds and along the same lines of thoughts as de-
43 4. Inspection of the model’s performance compared
scribed in [21] the OCR = f (p, q, e, θ) is defined for general
44 with laboratory tests
45 3D states by:
46 pei
ei0 − ex
The model’s performance is in this section evaluated
47 OCR = + , pxei = exp (22) by simulating experimental results for two different clays:
pei λ
48 335 Kaolin Clay from [3] and Lower Rhine Clay (experiments
49 conducted by the authors and Wichtmann and not pub-
||r||
50 lished yet). The low plasticity Kaolin Clay used for the ex-
e+ = A1/nf ei , A = 1 − . (23)
51 ||rb0 || periments has a liquid limit of wL = 47.2 %, a plastic limit
52 of IP = 12.2 % and a grain density of ρs = 2.590 g/cm .
3
53
295 Hence, OCR is a function of the void ratio e and the stress
54 invariants p, q and the Lode angle θ accounting for mul-340 The highly plastic Lower Rhine Clay stems from the ex-
tidimensional loading paths. Thus, one parameter less is avation side of an opencast lignite mine and has a liquid
55
56 required compared to the reference model [2] as the con- limit of wL = 56.1 %, a plastic limit of IP = 34.1 % and a
3
57 troversial parameter nocr introduced in [2] is omitted in grain density of ρs = 2.675 g/cm [3]. The permeability of
58
300 Eq. (22). Kaolin and Lower Rhine Clay has been determined by the
59 345 authors in a permeability test in a cylinder with variable
60 pressure head during the test resulting to k = 1.3 · 10−9
61 5
62
63
64
65
1 Table 3: Material parameters
2
3
Description Approx. range Kaolin Lower Rhine Clay
4
5 Transversal (hypo)elasticity
6 λ compression index 10−6 − 1 0.13 0.26
7 κ swelling index 10−7 − 0.6 0.05 0.04
8 νh Poisson ratio 0 − 0.5 0.3 0.2
9
α anisotropic coefficient 0−5 1.8 0.7
10
11 Intergranular strain (IS):
12 mR stiffness factor 1−7 3 5
13 R IS yield surface radius 10−6 − 10−2 2 × 10−4 2 × 10−4
14
d IS bounding surface radius 1 − 10 3 4
15
β0 IS hardening parameter 0−2 0.076 0.076
16
17 χ0 min IS exponent 1 − 10 7 7
18 χmax max IS exponent 20 − 50 40 35
19 Ca accumulation rate factor 0 − 0.2 0.001 0.005
20 Critical and loading surface:
21
Mc CS slope 0.5 − 2 1.0 0.95
22
23 ei0 max void ratio at pref = 1 kPa 0.5 − 5 1.76 2.47
24
25 fb0 loading surface factor 1−3 1.5 1.45
26 Viscosity:
27
Iv viscosity index 0−2 0.015 0.025
28
29
30
31 m/s and k = 3.6 · 10−11 m/s, respectively. Both materials The simulated effective stress paths (red) and the exper-
32 have been tested in disturbed form, i.e. using reconstituted imental results (blue) are given in Fig. 3 for tests C02,
33 samples. The material parameters used for the simulations C04 − C08. It is evident that the effective stress path
34
350 are listed in Tab. 3. 375 before cyclic mobility is significantly inclined towards the
35 left upper corner of the p − q plane [3]. This observation
36 4.1. Simulations with Kaolin Clay is reproduced very well with the AVISA model through
37 In several test series performed by Wichtmann & Tri- the anisotropic coefficient of α = 1.8 resulting in Eh =
38 antafyllidis [3] the stress amplitude q0 , the displacement 3.24 · Ev . The reference model could not reproduce the de-
39 rate ṡ, the initial mean pressure p0 , the OCR, the initial380 scribed slope of the effective stress path [2]. The inherent
40 anisotropy effects also the axial strain path, whereby the
355 stress ratio η0 = q0 /p0 and the cutting direction of the
41 sample first accumulates in extension. Due to the p − q in-
samples have been varied as described in [3, 10]. Each of
42 clination, first the compression triaxial critical state line is
43 these tests has been simulated with the AVISA model and
showed a good agreement with the experimental results reached, thus the accumulation of the axial strain changes
44
(see the following sections). Only one material parameter385 to compression as demonstrated in Fig. 4. The AVISA
45
set listed in Tab. 3 has been used for all simulations. The model captures well all of these features of the material
46
360
notation of the tests with CXX e.g. C02 is the same as in behaviour, whereas the reference model accumulated only
47
48 [3], so that missing information can be depicted from [3]. in extension [2]. For a performance based design of an e.g.
49 flooded opencast mine unter earthquake loading the accu-
50 4.1.1. Variation of deviatoric stress amplitude qampl 390 rate description of these effects is indispensable.
51 The tests C01 − C08 were performed with different de- The tests C05 and C06 have been performed with the same
52
365 viatoric stress amplitudes qampl = {30, 40, 45, 50, 60, 70} deviatoric amplitude qampl = 50 kPa but with different
53 kPa. The initial conditions OCR0 = 1, p0 = 200 kPa, initial void ratios e0 = 1.145 and e0 = 1.252, respec-
54 η0 = 0, ṡ = 0.1 mm/min were the same for all sam- tively. As reported in [3] lower initial void ratios due to
55 ples. In the simulations the same number of cycles as395 higher overconsolidation ratios lead to a higher number
56 documented in [3] has been applied. For the initializa- of cycles to reach cyclic mobility or the failure criterion
57
370 tion of the OCR the simulations used the initial void ratio of |ε1 | = 10 % used in the experiments. The simulations
58 are of course completely in agreement with the described
e0 = ei0 − λ ln(p0 /(1 kPa) OCR).
59 behaviour, presented also in Fig. 5 a) in terms of accu-
60
61 6
62
63
64
65
a) to sand whereas the reversal starts with the highest con-
1
2 1500 tractancy. This can present a reason why most clays do
Deviatoric stress q [kPa]
32
33 80 80
34
40 40
35
36 0 0
37
38 −40 −40
39 −80 −80
40
41 −6 −2 2 6 10 −6 −2 2 6 10
42 Axial strain ε1 [%] Axial strain ε1 [%]
43
44
45 Figure 4: Deviatoric stress vs. axial strain in undrained cyclic tests with different stress amplitudes (OCR0 = 1, p0 = 200 kPa, η0 = 0, ṡ = 0.1
46 mm/min). (simulation=red, experiment=blue)
47
48
49 the mean pressure leads to an acceleration of the accumu- ditions were chosen to p0 = 100 kPa, qampl = 30 kPa and
50 lated pore water pressure and of the residual strain. This ṡ = 0.1 mm/min.
51 behaviour is also captured with the model simulations il- The expected dilative response in the first cycle of undrained
485
52 lustrated in Figs. 9 and 10. 495 cyclic loading of overconsolidated samples is recognisable
53 The mismatch in the first cycle of the effective stress paths in the negative accumulated pore water pressure at N = 1
54 can be attributed to the fluctuation of the experimental in Fig. 11 a). Afterwards the material response is contrac-
55 initial void ratio. tive ṗ < 0 ⇒ u̇acc > 0. The number of cycles to failure
56 Nf = {116, 531} grows significantly with the OCR0 =
57 4.1.4. Variation of initial overconsolidation ratio OCR0 500 {1.5, 2.5}. The rate of εacc
1 and uacc decreases with OCR0
58
490 The influence of the initial overconsolidation ratio was resulting in a stiffer response of the material for higher
59 studied in 2 tests with OCR0 = {1.5, 2}. The initial con- OCR0 . All these effects are properly described by the
60
61 8
62
63
64
65
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16 Figure 5: Undrained cyclic tests with different stress amplitudes (OCR0 = 1, p0 = 200 kPa, η0 = 0, ṡ = 0.1 mm/min): a) normalized
accumulated pore water pressure uacc /p0 as a function of the number of cycles, b) accumulated axial strain as a function of the number of
17 cycles. (simulation=solid lines, experiment=dashed lines)
18
19
20 a) C09: ṡ = 0.01 mm/min b) C11: ṡ = 0.05 mm/min c) C12: ṡ = 0.5 mm/min
21
Deviatoric stress q [kPa]
34 80 80 80
35
36 40 40 40
37
0 0 0
38
39 −40 −40 −40
40
41 −80 −80 −80
42 0 40 80 120 160 200 0 40 80 120 160 200 0 40 80 120 160 200
43 Mean stress p [kPa] Mean stress p [kPa] Mean stress p [kPa]
44
45
46 Figure 6: Effective stress paths in undrained cyclic tests with different displacement rates (OCR0 = 1, p0 = 200 kPa, η0 = 0, qampl = 45
47 kPa). (simulation=red, experiment=blue)
48
49
50 models simulations, solid lines in Fig. 11. The effective stress paths are presented in Fig. 12 a) −
51 c), whereby the slope and the relaxation of the mean pres-
52 4.1.5. Variation of initial stress ratio η0 sure are properly reproduced with the model. Both, the
53
505 Normal consolidated samples with initial mean pres-515 experiments and the simulations, showed that the failure
54 sure of p = 200 kPa has been anisotropically consolidated criterion is reached due to excessive axial strain accumula-
55 to different deviatoric stresses leading to various initial tion and not due to large strain amplitudes as observed for
56 stress ratios in compression η0 = {0.375, 0.25, 0.125} and isotropically consolidated samples (compare Fig. 4 with
57 in extension η0 = {−0.125, −0.25, −0.5}. Then cyclic Fig. 13). Furthermore, the accumulation of the pore wa-
58
510 undrained shearing has been performed with qampl = 30520 ter pressure almost stabilized after a certain number of
59 kPa and displacement rate of ṡ = 0.1 mm/min. cycles and consequently a complete relaxation of p is not
60
61 9
62
63
64
65
1
2 a) C13: ṡ = 0.01 mm/min b) C14: ṡ = 0.5 mm/min
Deviatoric stress q [kPa]
1
10 [6] J. Graham, G. Houlsby, Anisotropic elasticity of a natural clay,
11 Géotechnique 33 (2) (1983) 165–180.
visco-elastic visco- [7] A. Amorosi, S. Rampello, An experimental investigation
12 hypoplastic 820 into the mechanical behaviour of a structured stiff clay,
13 Géotechnique 57 (2) (2007) 153–166.
14 [8] J. Duncan, H. Seed, Anisotropy and stress reorientation in clay,
15 transition J Soil Mech Found Div ASCE 92 (SM5) (1966) 21–52.
16 [9] A. Niemunis, Extended hypoplastic models for soils, habilita-
0
825 tion, monografia 34, Ruhr-University Bochum (2003).
17 10-6 10-5 10-4 10-3 10-2 [10] T. Wichtmann, Soil behaviour under cyclic loading - experimen-
18 [-] tal observations, constitutive description and applications, Verf-
19 fentlichungen des Instituts fr Bodenmechanik und Felsmechanik
20 am KIT. Habilitation. Heft 181. (2016).
21 830 [11] T. Wichtmann, K. Andersen, M. Sjursen, T. Berre, Cyclic tests
on high-quality undisturbed block samples of soft marine norwe-
22 gian clay, Canadian Geotechnical Journal 50 (4) (2013) 400–412.
23 Figure B.21: Shear modulus degradation curve and the definition of [12] C. Gu, J. Wang, Y. Cai, Z. Yang, Y. Gao, Undrained cyclic
R [1] and d.
24 triaxial behavior of saturated soft clay under variable confining
25 835 pressure, Soil Dyn Earthq Eng 40 (2012) 118–128.
26 [13] M. Zergoun, Y. Vaid, Effective stress response of clay to
775 • the critical and loading surface introduce 1 and undrained cyclic loading, Can Geotech J 38 (1994) 714–727.
27 2 additional material parameters, respectively. The [14] N. Duong, M. Suzuki, N. Hai, Rate and acceleration effects on
28
critical state slope in triaxial compression Mc can residual strength of kaolin and kaolin-bentonite mixtures in ring
29 shearing, Soils and Foundations 58 (2018) 1153–1172.
be calibrated through a triaxial test with samples840
30 [15] A. Niemunis, C. Grandas, L. Prada, Anisotropic Visco-
31 sheared up to high strains ensuring that the critical hypoplasticity, Acta Geotechnica 4 (4) (2009) 293–314.
780
32 state (e = ec and η = ηc ) is reached. The maximum [16] A. Whittle, M. Kiavvadas, Formulation of MIT-E3 constitutive
void ratio ei0 corresponds to the maximum void ra- model for overconsolidated clays, Journal of the Geotechnical
33 Engineering Division ASCE 120 (1) (1994) 173–198.
34 tio at the reference pressure pref = 1 kPa at the845 [17] G. Grimstad, S. Abate Degago, N. S, K. M, Modeling creep and
35 virgin isotropic consolidated line OCR = 1. For its rate effects in structured anisotropic soft clays, Acta Geotech-
36 calibration the most suitable test is an isotropic com- nica 5 (2010) 69–81.
pression test with constant rate of strain. In absence [18] S. Raude, F. Laigle, R. Giot, R. Fernandes, A unified thermo-
37
785
850 plastic/viscoplastic constitutive model for geomaterials, Acta
38 of this test, an oedometer test can be used instead as Geotechnica 11 (2016) 849 – 869.
39 explained in [2, 21]. The loading surface factor fb0 [19] N. Sivasithamparam, M. Karstunen, P. Bonnier, Modelling
40 controls the peak strength of the overconsolidated creep behaviour of anisotropic soft soils, Computers and
41 soil. Thus, it can be calibrated using triaxial data of Geotechnics 69 (2015) 46 – 57.
42 855 [20] T. Schanz, P. Vermeer, P. Bonnier, The hardening soil model:
790 initially overconsolidated samples. formulation and verification., in: Beyond 2000 in Computa-
43 tional Geotechnics. Ten Years of PLAXIS, A.A. Balkema, Am-
44 • the viscosity requires the calibration of only one sterdam, the Netherlands, 1999, pp. 281–290.
45 parameter denoted as viscosity index Iv . It describes [21] M. Tafili, T. Triantafyllidis, A simple hypoplastic model with
46 the intensity of creep for normalconsolidated states860 loading surface accounting for viscous and fabric effects of clays,
47 Submitted for IJNAMG.
and can be determined through a creep stage or con- [22] A. Niemunis, I. Herle, Hypoplastic model for cohesionless soils
48
795 solidation process evaluating the secondary settle- with elastic strain range, Mech. Cohesive-frict. Mater. 2 (4)
49 ment curve. Other methods as the consideration of (1997) 279–299.
50 two different isotachs can be used as well [9, 21, 2,865 [23] W. Fuentes, M. Hadzibeti, T. Triantafyllidis, Constitutive
51 model for clays under the isa framework, Vol. 80 of Holistic
23]. Simulation of Geotechnical Installation Processes - Benchmarks
52
and Simulations, Springer International Publishing, Germany,
53 2015, pp. 115–130.
54 References 870 [24] M. Tafili, T. Triantafyllidis, Constitutive model for viscous
55 clays under the isa framework, Vol. 82 of Holistic Simula-
800 [1] W. Fuentes, T. Triantafyllidis, ISA model: A constitutive model tion of Geotechnical Installation Processes - Theoretical Results
56 for soils with yield surface in the intergranular strain space, Int
57 and Applications, Springer International Publishing, Germany,
J. Numer. Anal. Meth. Geomech. 39 (11) (2015) 1235–1254. 2017, pp. 324–340.
58 [2] W. Fuentes, M. Tafili, T. Triantafyllidis, An isa-plasticity-based875 [25] M. Tafili, T. Triantafyllidis, On constitutive modelling of
59 model for viscous and non-viscous clays, Acta Geotechnica anisotropic viscous and non-viscous soft soils, Vol. 1 of Numer-
60
805 13 (2) (2017) 367–386.
61 18
62
63
64
65
1 ical Methods in Geotechnical Engineering IX, Taylor & Francis
Group, Porto, Portugal, 2018, pp. 139–148.
2 [26] Y. Dafalias, Bounding surface plasticity. I: Mathematical foun-
3
880 dation and hypoplasticity, J. Eng. Mech. 112 (9) (1986) 966–987.
4 [27] Y. Dafalias, M. Manzari, Simple plasticity sand model account-
5 ing for fabric change effects, J. Eng. Mech. 130 (6) (2004) 622–
6 634.
[28] M. Poblete, W. Fuentes, T. Triantafyllidis, On the constitutive
7
885 modeling of multidimensional cyclic loading, Acta Geotechnica
8 11 (6) (2016) 1263–1285.
9 [29] S. Krieg, Viskoses bodenverhalten von mudden, seeton und
10 klei, Veröffentlichungen des Instituts für Bodenmechanik und
11 Felsmechanik am KIT. PhD Thesis. Heft 150. (german) (2000).
890 [30] A. Niemunis, S. Krieg, Viscous behaviour of soils under oedo-
12 metric conditions, Can. Geot. &Journal 33 (1996) 159–168.
13 [31] J. Atkinson, Peak strength of overconsolidated clays,
14 Géotechnique 57 (2) (2007) 127–135.
15 [32] A. Niemunis, C. Grandas-Tavera, T. Wichtmann, Peak stress
16
895 obliquity in drained and undrained sands. smulations with neo-
hypoplasticity, Vol. 80 of Holistic Simulation of Geotechnical
17 Installation Processes - Benchmarks and Simulations, Springer
18 International Publishing, Germany, 2015, pp. 85–114.
19 [33] M. Hvorslev, Physical components of the shear strength of sat-
20
900 urated clays, ASCE Research Conference, Shear Strength of
21 Cohesive Soils, 1960, boulder Colorado.
[34] L. Guo, J. Wang, Y. Cai, H. Lui, Y. Gao, H. Sun, Undrained de-
22 formation behavior of saturated soft clay under long-term cyclic
23 loading, Soil Dyn Earthq Eng 50 (2013) 28–37.
24
905 [35] M. Hyodo, H. Tanimizu, N. Yasufuku, H. Murata, Undrained
25 cyclic and monotonic triaxial behaviour of saturated loose sand,
26 Soils and Foundations 34 (1) (1994) 19–32.
[36] Y. Vaid, J. Chern, Effect of static shear on resistence to lique-
27
faction, Soils and Foundations 23 (1) (1983) 47–60.
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61 19
62
63
64
65