See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/234377071

Numerical study of mono-strand anchorage mechanism under service load

Article  in  Structural Engineering & Mechanics · October 2004

DOI: 10.12989/sem.2004.18.4.475

CITATIONS READS

2 436

3 authors:

Daniel Marceau Mario Fafard

University of Québec in Chicoutimi Laval University
41 PUBLICATIONS   292 CITATIONS    273 PUBLICATIONS   2,022 CITATIONS   

SEE PROFILE SEE PROFILE

J. Bastien
Laval University
28 PUBLICATIONS   300 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

In situ velocity measurement of molten aluminum in the aluminum electrolysis cell View project

Non-destructive quality control of carbon anodes using modal analysis, acousto-ultrasonic and latent variable methods View project

All content following this page was uploaded by Mario Fafard on 25 March 2015.

The user has requested enhancement of the downloaded file.

Structural Engineering and Mechanics, Vol. 18, No. 4 (2004) 000-000 1

Numerical study of mono-strand anchorage

mechanism under service load
D. Marceau†
Department of Applied Sciences, University of Quebec at Chicoutimi, Quebec, Canada, G7H 2B1

M. Fafard† and J. Bastien†

Department of Civil Engineering, Laval University, Quebec, Canada, G1K 7P4

(Received , 2004, Accepted , 2004)

Abstract. Anchorage devices play an important role in post-tensioned bridge structures since they must
sustain heavy loads in order to permit the transfer of the prestressing force to the structure. In external
prestressing, the situation is even more critical since the anchorage mechanisms, with the deviators, are
the only links between the structure and the tendons throughout the service life of the structure. The
behaviour of anchorage devise may be studied by using the finite element method. To do so, each
component of the anchorage must be adequately represented in order for the model to approximate the
anchor mechanism as accurately as possible. In particular, the modelling of the jaw/tendon device may be
carried out using the real geometry of these two components with an appropriate constitutive contact law
or by replacing these components by a single equivalent. This paper presents the numerical study of a
mono-strand anchorage device. The results of a comparison between two different representations of the
jaw/tendon device, either as two distinct components or as a single equivalent, will be examined. In the
double-component setup, the influence of the wedge configuration composing the jaw, and the influence
of lubrication of the anchor, will be assessed.
Key words: anchorage device; contact; finite element; jaw/tendon device; lubrication; yielding; stress
and strain distribution; wedge penetration.

1. Introduction

Post-tensioning is a commonly used technique in the construction industry either for new
structures or to reinforce existing ones. In bridge design, post-tensioning regained popularity during
the 80’s with the introduction of new high performance concrete and the evolution of the post-
tensioning technique itself (material, methodology,...). This technique requires the tensioning of
prestressing tendons which are anchored to the concrete section using metallic anchorage devices.
Different types of the device of specific shapes and mechanical characteristics may be found on the
market. These are developed mainly by private companies involved in the prestressing industry and
are offered in a variety of capacities, according to the size and number of tendons to be anchored.

† Professor
2 D. Marceau, M. Fafard and J. Bastien

In this paper, the numerical behavior of a mono-strand anchorage device will be examined. In
particular, the influence of the anchor gripping action will be assessed.
The design and development of anchorage mechanisms is generally carried out by companies
specialized in the prestressing field using laboratory tests, and more recently, finite element analysis
(Bastien et al. 2002). To date, very few papers have been published involving the numerical analysis
of anchoring mechanisms. In this regard, Marceau (2001) and Bastien et al. (2002) represent one of
the few papers dealing with the numerical behavior of multi-strand anchorage mechanisms.
The finite element method figures as a complement only to laboratory tests because these tests,
while well adapted to inform about the global behaviour of the mechanism, cannot assess the state
of stress/strain developed in the core of mechanism parts. It should be made clear here that a typical
anchorage consists of a thick metallic cylindrical plate called the anchor head, which is perforated
by one or more conical cavities in which a tendon is gripped by a jaw. This jaw is made according
to the requirements of the specific post-tensioning system used, by two or three wedges called
herein split-wedges. The anchor head is set on the concrete section by means of another plate called
“anchor plate”, the latter being bolted to the concrete.
In order to ensure adequate behaviour representation, the interaction between the components of
the mechanism, as well as their non-linear behaviour, must be accounted for in a finite element
model. A proper model should, therefore, take into consideration the modelisation of : solids using
an elasto-plastic law and a large strain approach, contact between the anchor head and anchor plate
interface, contact between each jaw and the corresponding anchor head conical cavity and contact
between the jaw and the gripped tendon.
Marceau et al. (2002) and Marceau et al. (2001b) describe the numerical model that will be used
from now in the text, as well as the description of the different anchor components and the
associated contact interfaces. More over, the authors introduce a new constitutive contact law called
the gripping/contact law based on a phenomenological approach to simulate the jaw gripping action
on a tendon. Identification of the law parameters was done using the finite element method in
combination with a neural network approach and results of experimental tests conducted on five
mono-strand anchorage devices. This gripping/contact law will be used in this paper to study the
behaviour of a mono-strand anchor head in relation with the uniformity of the gripping action. Figs. 1
to 3 give geometric details of a jaw made of three metallic pieces assembled together (split-wedges)
at the top using a metallic ring. The finite element analysis which have been conducted to date were
all implicitly based on the fact that the spilt-wedges were uniformly distributed around the tendon

Fig. 1 Mono-strand anchorage

 Numerical study of mono-strand anchorage mechanism under service load 3

Fig. 2 Geometrical description of the anchorage device

and installed as such during construction, as shown in Fig. 4 for case 1. However, due to
construction practice, this condition represents an ideal situation, and wedge regrouping, such as
described in cases 2 and 3 of Fig. 4 are possible; case 3 being considered as the extreme situation.
Therefore, the aim of this paper is to describe the influence of different split-wedges and will
include configurations of mono-strand anchor head behaviour. Three different cases are examined
and numerical information regarding the von Mises equivalent stress distribution in the anchor head,
as well as at the surface periphery of the anchor head are presented. The penetration of the split-
wedges along the anchor head conical cavity is also presented under service loading. In order to
assess another extreme case in connection with the penetration of the jaw along the anchor cavity,
the practice of lubrication will be examined. Finally, the paper will provide additional information
related to the way the jaw-tendon assembly is represented in the numerical model.
The jaw and tendon may be represented as two different components or by a single equivalent
system. In the latter representation, the jaw/tendon gripping action is considered to be such that the
two components (jaw and tendon) are assumed to act as a unit. Therefore, the jaw/tendon assembly
is replaced by a single truncated conical component. This component has the same global
dimensions as the jaw/tendon assembly, and is meshed as such. Since, in practice, the wedges are
4 D. Marceau, M. Fafard and J. Bastien

Fig. 3 Details of the wedge gripping devices

thick pieces submitted to a surface hardening process, and since a tendon exhibits linear elastic
behaviour under service loads, the cone associated to the device is considered to be linear elastic.
For this jaw-tendon representation, the load is applied at the top of the truncated cone by imposing
a uniform pressure on the equivalent jaw/tendon system. In the case where the jaw and tendon are
considered as two different components, loading is applied by pulling the bottom end of the tendon.

2. Numerical model

The complexity of the mechanical behaviour of an anchorage device provides a broad range of
investigative possibilities. During the process of submitting the anchorage device to heavy
prestressing loads, a number of phenomena may be observed including : wedge-tendon gripping
 Numerical study of mono-strand anchorage mechanism under service load 5

Fig. 4 Details of the split-wedge configurations

action; friction at the wedge/cavity interfaces; friction the anchor block/anchor plate interface;
plastic strains; and large relative displacements at certain interfaces. An adequate finite element
model should be able to take into account all these mechanical phenomena.
This study, takes in to account three types of nonlinearity : large strain (large displacement),
plasticity, and contact between certain interfaces. In the latter, large relative displacement between
interfaces has been considered using a slave-master concept (Marceau 2001, Laursen and Simo
1993). The penalty/augmented Lagrangian method has been used for the evaluation of the stresses at
the various interfaces. This approach yields accurate results and makes non-compatible meshes
possible between different components of the anchorage device. All mathematical and numerical
techniques used for this model can be found in Marceau (2001). The model has been tested and
validated using information from the literature.
It should be noted in, that for the wedge-tendon contact interface (see Fig. 3), a new contact law
taking into account all the experimental observations is necessary (Marceau et al. 2002). Firstly, the
6 D. Marceau, M. Fafard and J. Bastien

Fig. 5 Normal stress law for the wedge-tendon interface

interface behaviour is considered orthotropic in the RTL directions. The normal contact stresses are
assumed to act in the radial direction (R), while the longitudinal direction (L) is considered parallel
to the tendon profile. While a Coulomb friction law is considered in the tangential direction (T), no
relative displacement is assumed between the tendon and wedges in the longitudinal direction (L).
Indeed, in this direction, and contrary to the wedge-alveolus interface, the relative displacement is
considered very small, allowing for the present simplification. In the radial direction, hyperbolic law
is proposed and defined as (see Fig. 5):
ηg
t R = ---------u (1)
g
where η represents the stiffness in the radial direction and g, the actual radial distance between the
wedge and the tendon, such that:

g = g0 – gu (2)

g u = – n ⋅ [ u ( X , t ) – u ( Ψ 0 ( ξ ), t ) ]
1 1 2
(3)
˜ ˜ ˜ ˜ ˜ ˜
1 2
where g0 defines the fictitious gap between the wedge and the tendon. The terms u and u
represent the displacement field of the slave and master contact surfaces respectively. ˜ ˜
The parameter g0 can be associated with the depth of the notches. As seen in Fig. 5, for a given η

ç
value, when g = g0 the radial stress is zero. At this point, the notches begin to penetrate into the
tendon and the stiffness increases progressively. When gu g0, the latter being the maximum
allowed penetration, the radial stress increases rapidly. This radial stress will increase more or less
rapidly depending on the value of η. Thus η and g0 can be considered as unknown parameters that
need to be identified.
 Numerical study of mono-strand anchorage mechanism under service load 7

From a mathematical point of view, the admissible conditions can be defined by:

g≥0 
g ≤ g0 

tR ≥ 0  (4)


t R ( X , t )g ( X , t ) ≥ 0 
1 1

˜ ˜
For the radial direction and

Φ = t T T – µ wt t R ≤ 0

·T t TT 
ξ – ζ ----------
1·
- = ---- tT 
t TT εT T 

·L 1  (5)
ξ = ---- t·TL 
εL 
ζ≥0 

Φζ = 0 
·L ·T
for the longitudinal and tangential directions with µwt, ξ and ξ , defining the Coulomb friction
coefficient in the tangential direction and the wedge-tendon relative velocities in the corresponding
directions. In the same manner, t·T L and t·TT represent the stress rate along the longitudinal and
tangential directions respectively. Regularization is obtained using the penalty numbers εL and εT.

3. Finite element simulation

Taking into consideration the geometrical symmetry of the mono-strand anchorage device, only
one-sixth of it has been represented when the split-wedges where considered uniformly distributed
around the tendon, as in case 1 (see Fig. 4(a)). When two or more split-wedges are grouped around
the tendon, as for cases 2 and 3, one-half of the anchorage device has been modelled (Fig. 4(b) and
Fig. 4(c)). The different meshes are presented in Fig. 6. For these meshes, symmetry conditions
have been applied in a standard way, while quadratic interpolation elements were used to discretize
the various anchorage components. These hexahedral and triangular prism elements were based on
the large strain deformation theory, and 3 × 3 × 3 Gauss points and 3 × 3 Hammer-Gauss points
were the schemes used respectively for integration. For the contact interfaces, eight-noded
quadrilateral and six-noded triangular elements were used. Their integration schemes were based on
a 3 × 3 Gauss points and a three Hammer points respectively. The numerical resolution was
performed using the Newton-Raphson and arc-length methods. Contact interfaces were solved using
a penalty coefficient of 105 in both normal and tangential directions, using four Lagrangian
augmentations.
Since each split-wedge occupies a conical portion of 112o inside the anchor cavity, in the ideal
case of their proper distribution inside this cavity, each split-wedge will be separated fromthe two
others at an angle of 8o. This angle corresponds to an arc length of about 2.4 mm at the top portion
8 D. Marceau, M. Fafard and J. Bastien

Fig. 6 Mesh of the anchorage device for each case

of the anchor cavity perimeter. Fig. 6(a) shows one-sixth of the discretized anchorage device, which
incorporates one-half wedge (sector of 56o). Therefore, a free space, corresponding to an angle of
4o, is apparent in Fig. 6(a). Along the two symmetric axes shown, usual symmetric conditions were
applied leaving one side of the wedge portion represented in Fig. 6(a) free from any boundary
conditions. It is also obvious in this figure that the jaw/tendon system exceeds the top of the anchor
head by 5.0 mm before loading, as observed in laboratory during testing. The anchor head was
 Numerical study of mono-strand anchorage mechanism under service load 9

considered to set on a rigid support in order to represent the experimental setup. The finite element
mesh associated with Fig. 6(a) is made up 3,828 elements, 4,057 nodes, and 12,171 degrees of
freedom.
As shown in Fig. 4(b), case 2 deals with the situation where two of the three wedges grouped side
by side, leaving two free portions of about 3.6 mm in arc length. For case 3, the three wedges are
side by side, leaving only one 7.2 mm arc length opening (see Fig. 4(c)). In these two situations,
one-half of the anchorage device (sector of 180o) has been discretized. As shown in Figs. 6(b) and
6(c), these meshes incorporate one entire wedge (sector of 112o) plus one-half wedge (sector of
56o). In these two situations, some additional contact interfaces arise from the wedge regrouping.
Indeed, one and two additional contact interfaces must be taken into account at the joining wedges
for cases 2 and 3 respectively. As can be seen in Fig. 6(c), a small gap still exists and appears at
one of two additional interfaces. This situation results from the complex geometry of the wedges, as
shown already in Fig. 3. In cases 2 and 3, 11,890 elements and 11,447 nodes constitute the meshing
leading to a total of 34,341 degrees of freedom for case 2 and 34,206 degrees of freedom for case 3.
Tables 1 and 2 show the anchor components and tendon mechanical characteristics used in the
numerical analysis. The anchor head exhibits an elasto-plastic behaviour while the split-wedges are
considered elastic, due to the high hardness of their surface corresponding to a heat treatment
process (Bastien 1992).
Consideration of contact is based on the work of Marceau (2001). The Coulomb frictional
coefficient between the anchor head and anchor plate (µhp) were set to 0.3 while the one between
the jaw and anchor cavity (µhw) was fixed to 0.09. As far as the gripping/contact law is concerned,
the following parameters were used: the stiffness parameter was set to η = 3,000 MPa, the Coulomb
frictional coefficient in the tangential direction to µwt = 0.3 and the maximum gripping depth to g0 =
1.5 mm as described in Marceau et al. (2002).

Table 1 Mechanical properties of the anchorage device

Yield limit Ultimate stress Strain at failure
Component
(MPa) (MPa) (%)
Anchor head 400 750 16
Anchor plate 260 400 22
Wedge 400 Elastic behaviour
5
Young modulus: 2,0 · 10 MPa, Poisson ratio: 0,30

Table 2 Characteristics of the tendon

Effective area (mm2) 139
Nominal stress at failure (MPa) 1770
Nominal force at failure (kN) 246
Admissible stress (MPa) (0,8 · 1770) 1416
Admissible stress (kN) 197
Young modulus: 2,0 ·105 MPa, Poisson ratio: 0,30
10 D. Marceau, M. Fafard and J. Bastien

4. Results and discussion

As prescribed by many design codes (Bastien et al. 2002), the admissible service load corresponds
to 80 percent of the ultimate strength of the tendon (0.80Fu). This load level represents the
maximum prestressing force allowed during jacking procedure. Considering the characteristics of
the tendon described in Table 2, all finite element results are presented for a maximum load of 196
kN, corresponding to 0.80Fu. The first part of this section deals with the influence of the jaw/tendon
device on the behaviour of the anchorage mechanism. Secondly, some results are presented
regarding the influence of the split-wedges configuration (cases 1 to 3) and finally, the numerical
model is used to simulate the effect of lubrication of the jaw/cavity interface.

4.1 Jaw/tendon device versus truncated cone

Fig. 7 shows the influence of the real jaw/tendon device on the behaviour of the anchorage
mechanism, comparatively to a monolithic truncated cone for the numerical simulation of anchorage
mechanisms, as performed by Marceau et al. (2001a). In the former, three dimensional analysis
were performed using the mesh shown in Fig. 6(a). For the latter, the particular symmetry condition
implies the possibility to solve the problem using axisymmetrical formulation. To provide a valid
comparison, both models were studied using the same Coulomb frictional coefficient (0.09) between
the jaw and the cavity. Firstly, Fig. 7(a) shows that the penetration of the jaw/tendon device is larger
than that occurring with the truncated cone (3.3 mm versus 0.4 mm at (0.80Fu)). The small
penetration of the cone is due to its large stiffness in the radial and circumferential directions. In
this sense, it is clear that the truncated cone allows lower values of stress and strain that those

Fig. 7 Influence of the split-wedges on the behaviour of the anchorage device

 Numerical study of mono-strand anchorage mechanism under service load 11

obtained with the jaw/tendon device. For the jaw/tendon device, the gripping/contact constitutive
law, as well as the gap between wedges, allow progressive penetration of the triangular notches in
the tendon. It should be noted that similar values for the penetration of the jaw/tendon device in the
cavity were obtained from experimental tests (Marceau et al. 2002). This penetration of the jaw/
tendon device causes an increase of the yielded volume of the anchorage head, as shown in Fig. 7(b),
comparatively to the truncated cone. Considering these facts, it can be hazardous to draw practical
conclusions from the numerical simulations or tests using truncated cones.

4.2 Misalignment of the wedges in the cavity

In this sub-section, numerical simulations were performed to study the influence of splitwedge
misalignment on the mechanical behaviour of the anchorage mechanism. The three most critical
split-wedge configurations are described in Fig. 4, and defined as cases 1 to 3 respectively. As

Fig. 8 Axial strain for three levels along peripheral face of the anchor block
12 D. Marceau, M. Fafard and J. Bastien

Fig. 9 Hoop strain for three levels along peripheral face of the anchor block

described in Fig. 2, the total height of the anchor block is 55 mm. Figs. 8 to 10 show the strain
distribution on the peripheral face of the anchorage head at three different levels: 15, 30, and 45 mm
from base. Referring to Fig. 4 for the definition of the angle (θ ), these last figures shows that there
is a large difference between strains obtained with the split-wedge and the truncated cone, which is
a constant value. In particular, it is shown that strains increase for angles corresponding to the
middle of a wedge and, therefore, to the location between the wedge and the anchor block where
pressure value is maximum. In this manner, the strain values are higher at the top of the anchor
block and decrease to the bottom. Also, Fig. 10 shows that radial strain distribution causes the
anchor block to deform according to an elliptic shape. This situation has the effect of reducing the
total contact area between the tendon/wedge device and the cavity and, therefore, the efficiency of
the anchor block.
On the other hand, the Fig. 11 shows the von Mises stress distribution along the same paths. It
can be concluded that the misalignment of the split-wedge does not affect the results significantly.
 Numerical study of mono-strand anchorage mechanism under service load 13

Fig. 10 Radial strain for three levels along peripheral face of the anchor block

However, the use of the real jaw/tendon provides the possibility to identify yielded regions for
which the truncated cone has never reached the yield stress. Also, Fig. 11 shows that von Mises
stress increases from the region located between two wedges (θ = 0o, 100o) to a maximum value
located near to the middle of a wedge (θ = 60o, 180o). Along the height of the anchor block,
yielding appears from the top (45 mm) and spreads over the second level (30 mm), but not over the
level corresponding to 15 mm. Compared to the results obtained with the jaw/tendon device, it can
be concluded that those of the truncated cone are largely underestimated.
Finally, Fig. 12 shows that the misalignment of the split-wedge has no significant effect on the
penetration of the jaw/tendon device in the cavity or on the yielded volume. However, this is not the
case if these results are compared with those obtained with the truncated cone (see Fig. 7) for which
the difference is very significant.
14 D. Marceau, M. Fafard and J. Bastien

Fig. 11 von Mises stress for three levels along peripheral face of the anchor block

Fig. 12 Influence of the split-wedge configuration on the behaviour of the anchorage device
 Numerical study of mono-strand anchorage mechanism under service load 15

Fig. 13 Influence of lubrication on the behaviour of the anchorage device

4.3 Lubrication of the jaw/cavity interface

The lubrication of the jaw/cavity interface can be studied by reducing the Coulomb frictional
coefficient at this interface. To do so, the value is modified from 0.09 to 0.02, which can be
compared to a lubrication using graphite paste. Using the ideal split-wedge configuration (see Fig. 4),
Fig. 13 shows that lubrication has a significant effect on the penetration of the jaw/tendon device
and the yielded volume. Particularly, Fig. 13(b) shows that the whole volume of the anchor bloc is

Fig. 14 Influence of lubrication on the equivalent plastic strain distribution

16 D. Marceau, M. Fafard and J. Bastien

yielded for the lubricated case. In the same way, Fig. 14(b) indicates that the equivalent plastic
strain increases by about 130 percent for the lubricated case, compared with the case without
lubrication, which is in agreement with the excessivepenetration of the wedges as shown in Fig. 13(a).

5. Conclusions

In this paper, a comparative finite element study was presented to verify the influence of the jaw/
tendon device on the mechanical behaviour of an anchorage mechanism. To do so, simulations were
performed using a truncated cone, which can be seen as an idealization of the real jaw/tendon
device, and finally, with the real jaw/tendon device itself. The latter was modelled using particular
contact constitutive law at the jaw/tendon interface to adequately simulate the penetration of the
triangular notched into the tendon. The results show that such modelling provides a better way to
represent the real behaviour of the anchorage mechanism which has been confirmed by experimental
tests. In particular, the use of the jaw/tendon device has a significant effect on the yielding of the
anchor block, due to the excessive penetration of the jaw/tendon device into the alveolus. This
situation is mainly due to the gap between each wedge composing the jaw, providing the possibility
for each of them to be easily deformed along hoop and radial directions.
Also, simulations were carried out to verify the effect of misalignment of the split wedges
composing the jaw. The considered cases show that misalignment has no significant effect on the
penetration of the jaw/tendon device and yielding of the anchor block. However, the results
presenting strains and von Mises stress distributions demonstrate that the higher values are located
in the regions where contact pressures are maximum, such as in the middle of a wedge.
Finally, some results are presented regarding the effect of lubrication of the jaw/alveolus interface.
This procedure, which is very widespread in practice, is very faithful and is not recommended due
to the effect of the excessive penetration of the jaw/tendon device in the anchor block, and of
course, the high yielding level reached in the anchor block.

Acknowledgments

The authors would like to gratefully acknowledge the collaboration of Mr. Jean-Xavier Gartner,
engineering student at the catholic school for the Arts and Crafts, France (Ecole Catholique des Arts
et Métiers, France). Partial support of this research by the Quebec Ministry of Transportation, the
Natural Sciences and Engineering Research Council of Canada, and the Quebec Funding for
Research in to Nature and Technology (Fonds québécois de la recherche sur la nature et les
technologies), is also gratefully acknowledged.

References

Bastien, J. (1992), “Comportement des te^ tes d’ancrage mono et multi-torons: Étude expérimentale et étude par
éléments finis”, Ph.D. thesis, Département de génie civil, Université Laval, Québec.
Bastien, J., Marceau, D., Fafard, M. and Ganz, H. (2002), “Mechanical behaviour of multistrand anchorage
mechanisms under service and ultimate loads”, J. Bridge Eng., ASCE, Submitted.
 Numerical study of mono-strand anchorage mechanism under service load 17

Laursen, T.A. and Simo, J.C. (1993), “A continuum-based finite element formulation for the implicit solution of
multibody, large deformation frictional contact problem”, Int. J. Num. Meth. Engng., 36, 3451-3485.
Marceau, D. (2001), “Modélisation du contact tridimensionnel en grands déplacements relatifs et son application
à l’étude des dispositifs d’ancrages multitorons”, Ph.D. thesis, Département de génie civil, Université Laval,
Qu´ebec.
Marceau, D., Bastien, J., Fafard, M. and Chabert, A. (2001a), “Experimental and numerical studies of mono-
strand anchorage”, Int. J. Struct. Eng. Mech., 12(2), 119-134.
Marceau, D., Fafard, M. and Bastien, J. (2002), “Constitutive law for wedge-tendon gripping interface in
anchorage device – Numerical modeling and parameters identification”, Int. J. Struct. Eng. Mech., 15(6), 609-
628.
Marceau, D., Fafard, M., Bastien, J. and Boudjelal, M. (2001b), “Application des réseaux neuromimétiques à la
détermination du comportement déformationnel d’un dispositif d’ancrage multitorons”, In “Actes du
cinquième colloque en calcul des structures”, 1, 189-196. Ed. Batoz, Ben Dhi & Chauchot (Tecknea,
Toulouse), Giens, France.

View publication stats