International Journal of Sustainable Built Environment (2013) 2, 1–9

Gulf Organisation for Research and Development

International Journal of Sustainable Built Environment

ScienceDirect
www.sciencedirect.com

Prefabricated mud wall units based on a traditional Japanese

building system: Lateral in-plane performance in terms of
connection, and crack development using digital speckle photographyq
Hassel B. Ivón a,⇑,1, Kitamori Akihisa a,1, Kiho Jung b,2, Komatsu Kohei a,1
a
Kyoto University, Research Insititute for Sustainable Humanosphere (RISH), Laboratory of Structural Function, Gokasho, Uji, Kyoto 611-0011, Japan
b
Shizuoka University, Faculty of Education, 836, Ohya, Suruga-ku ,Shizuoka-Shi, Shizuoka-Ken 422-8529, Japan

Received 30 June 2012; accepted 1 October 2013

Abstract

A shear wall composed by prefabricated mud wall units (PMWU) was studied. Mud panels reinforced by bamboo laths are enclosed
in wooden frame, combining mud material’s high energy absorption and stiffness of wooden frame. Among the advantages are
workability and quality control before and after assembly. The PMWU’s performance related to the number of frames’ connectors
was analyzed. Digital speckle photography (DSP) was used to measure strains and finite element method (FEM) to validate the results.
The shear stiffness of PMWU varied due to rotation of the panels. The stress transmission was influenced by the number of connectors
via shear and contact between frames.
Ó 2014 The Gulf Organisation for Research and Development. Production and hosting by Elsevier B.V.
Open access under CC BY-NC-ND license.
Keywords: Prefabricated construction system; Mud wall; Eco material; Digital speckle photography; Timber frame

1. Introduction

Low-cost materials are currently in use for faster build-

⇑ Corresponding author. Present address: Department of Engineering ing construction; however, some modern building materials
Sciences, Applied Mechanics, The Ångström Laboratory, Lägerhyddsvä- can cause “sick building syndrome” (SBS) (MLIT, 2003).
gen 1, Box 534, SE-751 21, Uppsala, Sweden. Phone: +46 (0)18 4712790. Motivated by the desire to improve and “modernize” nat-
E-mail addresses: [email protected], [email protected] (B.I. Hassel),
[email protected] (J. Kiho). ural construction systems that are environmentally
1
Tel.: +81 774 38 3670; fax: +81 774 38 3678. friendly, we investigated tsuchikabe. Tsuchikabe (in Japa-
2
Tel.: +81 54 238 4677. nese), also known as mud wall, is a collective designation
q
Partial presentation of this article was presented at the Annual meeting for walls built with clay and mud using traditional Japa-
of the Japanese Wood Research Society, in Matsumoto, March 15-17, nese craft techniques. It is also known as wattle-and-daub
2009.
Peer-review under responsibility of The Gulf Organisation for Research
in Western countries, and has been used worldwide since
and Development. the Bronze Age.
Earth is among the simplest, cheapest, and most durable
building materials. It is a good thermal and acoustic
Production and hosting by Elsevier insulator, as well as flame, rot and insect resistant. Mud

2212-6090 Ó 2014 The Gulf Organisation for Research and Development. Production and hosting by Elsevier B.V. Open access under CC BY-NC-ND license.
http://dx.doi.org/10.1016/j.ijsbe.2013.10.002
2 B.I. Hassel et al. / International Journal of Sustainable Built Environment 2 (2013) 1–9

walls are of great importance in traditional Japanese archi-

tecture. They are used for town houses, temples and
shrines, particularly in Kyoto-style houses (Kyomachi-ya),
which are among the most traditional and ancient style
of housing in Japan. Their manufacture includes natural
materials and no chemicals, providing a healthy dwelling
environment (Saijo et al., 2004; Yokobayashi and Sato,
2007). The mud used in tsuchikabe is a mixture of clay,
earth, sand, and rice straw. Clay acts as a binder and holds
the mixture together. Aggregates like earth and sand pro-
vide bulk and dimensional stability. The mixture is allowed
to age over a period of months until maturation. When
used for wall construction, the straw remaining after this
process strengthens the wall (Shimase et al., 2007), controls
shrinkage, and provides long-term flexibility. Moreover, a
bamboo skeleton adds the necessary tensile and out-of-
plane strength.
The mud walls used in residential houses need to be thin
but still ensure high strength with energy absorption so as Fig. 1. Setup of a mud wall composed of prefabricated mud wall units
to withstand seismic forces. The infill mud material and the (PMWUs) linked to a frame by steel screws.
timber frame combine their independent strength and stiff-
ness properties, providing balance to the structure. mechanical properties. It is also of our interest to set a
Conventional mud wall construction involves high cost, starting point on the development of a non-contact method
skilled labor, a long hydration and drying process for the that can be used for the study of the structural state of
mix to harden (about 3 months at the construction site), existing mud wall buildings. Overall, our contribution is
and difficult to repair after setting. Hence, the use of tsuc- expected to be very useful for practical studies of mud walls
hikabe is decreasing, leading to a rapid disappearance of with different material properties and connection tech-
specialized skilled craftsmen. niques. It is overriding to understand the behavior of each
Tabuchi et al. (2006) tested large conventional mud wall component of the proposed structure, as well as developing
specimens (1960 by 2500 mm) in shear. The stiffness of computational model that can be used for further improve-
those specimens (using a mechanical model which includes ments or modifications.
the effect of the frame, joints, and mud material) was 591
KN/rad. Komatsu et al. (2008) developed a system inspired
by Japanese tradition, named the prefabricated mud wall 2. Materials and methods
unit (PMWU). The units are composed of mud panels rein-
forced by an inner perpendicular bamboo grid enclosed in Fifteen pre-fabricated mud wall units (PMWUs) of
a wooden frame. Designed with manageable dimensions of 50 mm in thickness and 450 mm on each remaining side
450 by 450 mm, PMWUs are produced and cured at the (Fig. 2) were produced and tested in shear. The units,
factory. Moreover, PMWUs could be stored for future wattle-and-daub-like panels, with a skeleton made of a
use and produced under controlled environments making perpendicular bamboo grid, were enclosed in a 50-mm-
them easier to handle, repair and replace even after instal- thick Japanese cedar (Cryptomeria japonica D. Don)
lation. Walls can be assembled in situ using steel screws to frame (Fig. 2c). The bamboo skeleton was loosely inserted
link the necessary number of PMWUs (Fig. 1). in the inner frame. Japanese mud was applied in two lay-
We aimed to verify the stress transmitting mechanism of ers: an inner layer called arakabetsuchi and an external
high lateral resistance of PMWUs by varying how wall layer called nakanuritsuchi. The inner layer was applied
panels are interconnected. Samples using different number in two steps on both sides. Arakabetsuchi was composed
of connectors linking the units to the external frame were of 100 liters of sandy clay and straw fibers (warasusa,
tested; mechanical properties such as stiffness and strength 0.4 kg). The external mud layer (nakanuritsuchi) was
were compared. The full strain field on the surface of the applied after an interval of one month. Nakanuritsuchi
mud material was studied through digital speckle photog- was composed of 100 l of sandy clay, straw fibers (wara-
raphy (DSP). Through DSP the in-plane strain fields were susa, 0.68 kg) and 40 l of sand (HOWTEC, 2004). The
obtained. Moreover, we report on the load named LoadF, sand prevents the mud from cracking when drying. The
defined as the load that produced the first apparent crack units were dried at room temperature over a period of
on the mud material. A finite element model (FEM) was 3 months.
created and compared with the experimental results. We For testing, the units were put inside a four-corner pin-
aim to develop a green, manageable construction system jointed outer frame of Spruce (Picea abies), and linked
based on an ancient Japanese system, with controllable using steel screws (4.8  90 mm) (Fig. 2b). The bottom
 B.I. Hassel et al. / International Journal of Sustainable Built Environment 2 (2013) 1–9 3

Fig. 2. Diagram of the PMWU setup. (a) Rear view of the specimen, showing displacement transducers. (b) Front view, prepared for digital speckle
photography (DSP) analysis, with a white speckle applied over a black background. (c) Manufacturing process of PMWU: bamboo grid inside the inner
frame covered by mud. Ch, channel (of displacement transducer).

member of the outer frame was fixed to the testing frame. the units. This is because the technique requires a recogniz-
Performance of the lateral resistance of the PMWUs was able pattern on the specimen in order to register displace-
studied under five linking conditions. Each type (N0-N4) ment (Fig. 2b). A spotlight was used to enhance the
was defined by the number of screws on each side which contrast between the speckle patterns and the background.
varied from zero to four screws. Three specimens of each
type were tested. A horizontal monotonic load was
applied through the upper member using an actuator at 2.1. Digital speckle photography (DSP)
a constant moving rate of 0.15 mm s 1 until failure
(Fig. 2a). The equipment used was a Riken oil pump unit We used digital speckle photography (DSP) and a com-
(Model MP-4ALS55) and a 50-kN load cell (Tokyo Sokki mercial non-contact measurement software Aramis 1.3 M
TCLM-50KNB). Displacement transducers (Tokyo Sokki by GOM to analyze the full strain field on the images. Pic-
CDP-50) were set on the back side of the specimens tures of the front side of the samples were taken every 5 s
(Fig. 2a). The deformation rate was controlled from the throughout the test with a Canon EOS KISS Digital X
displacement read from CH1. Using the perfect bi-linear camera. The applied load was synchronized with the cap-
approximation (Architecture Institute of Japan, 2002) on turing time of the pictures. DSP uses digital image correla-
the retrieved load-shear deformation angle curves, the tion algorithms to directly measure local displacements.
stiffness, yield strength, and energy absorption of the The area of interest was masked and subdivided into facets,
specimens were calculated. Channels CH6, CH7, CH8 which are partially overlapping areas. In this case the facets
and CH9 were used to calculate the relative rotation were 15 by 15 pixels with a 2-pixel overlap. This allowed a
of the panel units relative to the outer frame, spatial resolution of 13 pixels in the displacement measure-
hpanel = (CH6 + CH8)/h–(CH7 + CH9)/l, (see Fig. 2a for ment. Images were taken before and during displacement.
the definition of h and l). Facets were tracked by recognizing their gray-scale pattern
To perform digital image analysis, white speckles were with sub-pixel accuracy; cross-correlation was performed
airbrushed over a black background on the front side of to obtain the displacement field of the unstrained and
4 B.I. Hassel et al. / International Journal of Sustainable Built Environment 2 (2013) 1–9

strained stages. Finally, the displacement was differentiated dimensions as the actual PMWU. All the elements are
to obtain the strain field. four-node isoparametric with bilinear interpolation (class
4) (MSC Software Vol. B, 2005). The mud and wooden
2.2. Definition of characteristic properties by the bi-linear parts were modeled with type 3 elements (software’s
model and DSP nomenclature), which are a plane stress quadrilateral with
two degrees of freedom per node.
The corresponding parameters are illustrated in Fig. 3. A contact body is a set of curves, surfaces, or elements
Based on the bilinear model (Architectural Institute of that act as a body under a contact analysis (MSC Software
Japan), the initial stiffness (K) was calculated as the slope Vol. A, 2005). In this case, a contact body represents each
of the line connecting the origin point and the yield load component part of the test specimen (Fig. 4a). Nine
(Py). The yield load was determined by tracing a line that deformable contact bodies were defined, four per frame
crosses 0.4 Pmax and 0.9 Pmax and then shifting it until the and one for the mud panel including the vertical and hor-
line becomes tangent to the curve, the later traced in Fig. 3. izontal bamboo bars. The contact conditions and friction
In order to define the ultimate load (Pu) for the elasto-plas- coefficients were indicated in a contact table (MSC Soft-
tic curve, the area under it and the experimental load-slip ware Vol. C, 2005). For the contact between the mud body
curve (absorbed energy) up to 0.8 Pmax after the maximum and the inner frame, l = 0.90, (Fig. 4b, point A). For the
load (Pmax) should be equivalent. If the load capacity does external frame-to-unit contact, l = 0.2. These coefficients
not decrease until 0.8 Pmax after reaching Pmax, the equiv- of friction were set with the intention of simulating the
alent area is calculated as the value under the curve up to a experimental setup conditions. These included a “glued”
deformation angle of 1/15 rad. The apparent average crack (terminology used by the software) boundary condition
load (LoadF) was obtained from the curves of the number between the mud panel and the inner frame, and almost
of facets (from DSP) traced against the load level (Fig. 3, no friction between the unit and the outer frame during
curve on the right). Each digital image is subdivided into the elastic loading.
facets. In the elastic area of the load-deformation curve, The boundary conditions included fixed horizontal and
corresponding to the global experimental results, the num- vertical displacements for the bottom nodes and fixed ver-
ber of facets on the surface of the mud does not change. tical displacements for the top nodes of the outer frame
The number of facets starts decreasing below the PMWU’s (Fig. 4). The shear tests were simulated below the yield
yield point (Fig. 3). Considering that mud is a very brittle load. Links between nodes (beam-column) emulated the
material, no large deformations are present. The decrease pin-jointed corners of the external frame (Fig. 4b, point
in the number of facets coincided with the appearance of B), as well as the screws inserted between the outer frame
the first apparent crack. The load corresponding to this and the unit’s inner frame (Fig. 4b, point C). The respective
stage was defined as the failure load, LoadF. displacement and rotational restrictions were set to the
start and end nodes of each corresponding link (software
input).
2.3. Finite element model (FEM) As for the perpendicular bamboo grid, a four-node iso-
parametric rebar element (Marc Mentat type 143) was used
A two-dimensional model was created using the com- in conjunction with type 3 continuum elements (host ele-
mercial software Marc Mentat by MSC with the same ments). These are four-node hollow quadrilaterals in which
single strain members, such as reinforcing rods (that is,
rebars), can be placed (MSC Software Vol. B, 2005).
The rebar technique allows the use of separate constitu-
tive theories for the rebar and the host elements. Compat-
ibility between the rebar and the host element is enforced
by superimposing both using the same element connectiv-
ity. Therefore, the two elements share the same space and
element connectivity, so that no slip in the interface is
allowed. The element is integrated using a numerical
scheme based on a Gauss quadrature with two integration
points per rebar (MSC Software Vol. B, 2005). The rebars
were entered in intervals of 50 mm with an orientation cor-
responding to 0° and 90° relative to the frame. The degrees
Fig. 3. Load-deformation (left) and load-facets curves (right), correlated of freedom of the nodes in the inserted elements list are
by the load level. From the bilinear method, the elasto-plastic curve is automatically tied using the corresponding degrees of free-
overlaid, defining the initial stiffness (K), yield load (Py), and ultimate load
dom of the nodes in the host elements. The bamboo mate-
(Pu). The load crack (LoadF) was defined with the aid of DSP, plotting the
number of facets that compose the digital image of the specimen. The rial was defined as isotropic, Et = 19000 N mm 2 (tensile
decrease in number of facets indicates failure of the correlation between modulus) and m = 0.4 (Poisson ratio) (Laroque P, 2007).
consecutive stages. The mud was modeled as isotropic so that
 B.I. Hassel et al. / International Journal of Sustainable Built Environment 2 (2013) 1–9 5

Fig. 4. Setup and material orientation of the PMWU. (a) Component parts of the test specimen. (b) Boundary conditions of the FE model, showing each
body and the contact between them (A); Pin-jointed outer frame, emulated by links (B). Specimen type-N3, using three screws per side as inner-to-outer
frame ties, simulated by links between nodes (C). A horizontal load applied along the center line of the upper beam, P.

E = 200 N mm 2 and m = 0.167 using the results from

Nakao and Yamazaki (2006). The PMWU frame was made
of spruce (Picea abies) (q = 450 kg m 3 (density), ET/EL/
ER = 397/12800/625 N mm 2 (Young’s Modulus), GLR/
GRT/GLT = 617/71/587 N mm 2 (Shear Modulus) (Keunecke
et al., 2007b; Hassel et al., 2009) while the external frame
was made of Japanese cedar (Cryptomeria japonica D.
Don) (q = 389 kg m 3, ET/EL/ER = 22/6009/685 N mm 2,
GLR/GRT/GLT = 1061/32/894 N mm 2, obtained with the
SCA-method (Berard et al., 2006; Hassel et al., 2008,
2009)). Both the PMWU frame and outer frames were
modeled as orthotropic materials.
The load was applied at a vertical line of nodes along the
center of the upper member of the outer frame. A constant Fig. 5. Average experimental global load-shear deformation curves for
time stepping procedure was used to apply a 20-kN load in each PMWU type, displaying a similar global behavior for each type. N0–
twenty steps over a total loadcase time of 200 s. For the N4 indicates the number of screws (0–4) used per edge/side.
analysis the full Newton–Raphson solution control was
used. Table 1
Initial stiffness of prefabricated mud wall units obtained experimentally
(bilinear method) and predicted by finite element model (FEM).
3. Results and discussion
Stiffness [N/rad]

3.1. Shear behavior of PMWU in terms of the number of Sample Experimental FEM
Type-N0 170068 189076
connectors Type-N1 229213 250000
Type-N2 246568 254237
The average global experimental load-shear deforma- Type-N3 325544 276073
tion curves are plotted in Fig. 5. Their corresponding elas- Type-N4 276852 346021
tic stiffness values are displayed in Table 1, together with
their FEM predictions. The bilinear or perfect elasto-plas-
tic method was applied to these curves. The results shows. Specimen type-N3 presented the highest stiffness,
obtained from the analysis are summarized in Table 2, at the cost of having the lowest average energy absorption
combined with the DSP data. The values of the properties capacity. This can be explained by closely observing its ulti-
are averaged over the sets of each type of sample. It should mate load, Pu. Pu is defined as the load that produces an
be noted that the stiffness was particularly influenced by the equivalent area under the load-slip curve (absorbed energy)
number of screws, showing a tendency to improve in accor- as the 0.8 Pmax after Pmax. When the load-slip curve does
dance to the number of connectors up to type-N3. No not reach 0.8 Pmax, Pu is calculated for an ultimate defor-
improvement was observed beyond type-N3, as Fig. 6 mation of 1/15 rad. In the case of type-N3, 1/30 rad was
6 B.I. Hassel et al. / International Journal of Sustainable Built Environment 2 (2013) 1–9

Table 2
Summary of the results obtained from the bilinear method and digital speckle photography (DSP) at the first apparent crack stage.
Bi-linear method DSP method
Sample Number Max load Py Stiffness Energy Pu LoadF exF eyF cxyF Stiffness
of screws [kN] [kN] [kN/rad] [kN rad] [kN] [kN] [x 10-6] [x 10-6] [x 10 6] [kN/rad]
Type-N0 0 3.92 2.82 170.068 0.19 3.57 2.01 127 3911 2109 189.076
Type-N1 1 4.06 2.83 229.213 0.22 3.78 1.44 61 96 3126 250.000
Type-N2 2 4.13 2.57 246.568 0.22 3.78 1.76 1209 363 3811 254.237
Type-N3 3 4.84 3.12 325.544 0.15 4.42 2.38 42 873 3411 276.074
Type-N4 4 4.43 2.80 276.852 0.20 3.99 2.53 1257 67 4450 346.021

The yield load (Py) from the bilinear model, and LoadF
from DSP are plotted in Fig. 3. The number of facets
against deformation corresponds to the DSP results. This
curve is directly related to the global experimental load-
deformation data for a PMWU. The strains exF, eyF and
exyF in Table 2 are average values obtained from the mask-
ing of the speckles on the mud material when applying
LoadF. From Fig. 3, we observed that the structural behav-
ior of the PMWU was not affected by the appearance of the
first crack, since the elastic slope of the load-deformation
curve was also unaffected when we exceed LoadF. We
believe that this is due to the combination of the bamboo
Fig. 6. Comparison between global mechanical properties obtained
through the bilinear method. Each value corresponds to the average of
grid, mud and fibers contained in the mud panel. Therefore
three replicas of each type of PMWU. the mud itself displays a brittle behavior, while the
PMWUs overall have a ductile behavior, even after the sur-
face was cracked.
registered and therefore used for the calculation of Pu
(Fig. 5). Specimen type-N4 showed a similar performance,
although it implies less economy in terms of materials and 3.2. Separated effects of the PMWU’s components on the
handling time. total deformation
The stiffness of the mud panels obtained by digital
speckle photography shows the influence by the number Fig. 7 displays the rotation of the component parts of the
of connectors, with the stiffness increasing accordingly test specimen separately. Fig. 7a presents the DSP shear
(Table 2). These results agree with the experimental ones angle curves of the mud panels plotted against the load,
(Table 2), which also exhibit an improvement on the ulti- and represented by their corresponding approximation
mate load and a maximum load resisted by the structure, lines. These curves are average values from the central
given the strongest connection between inner and outer region of the mud panel. As can be noted by comparing
frames. Type-N0 behaved in an interesting way, resisting the curves in Fig. 7a, the initial shear stiffness does exhibit
relatively high load levels (LoadF). From the DSP results considerable variation among PMWU types. Fig. 7b, on
we assumed that the later effect is due to high compression the other hand, displays the relative rotation between the
on the corners of the mud panels when applying a shearing inner and outer frames, directly showing the influence of
force (Table 2, eyF and Fig. 10). the connectors. The high impact of the number of steel con-
It is apparent that the difference in the results was pro- nectors on the rotation of the panels is apparent. Lines cor-
duced by the specimen type-N4, with lower values of responding to the elasticity relative to rotation as obtained
Max load, Py, Pu and stiffness (Table 2). This can be by FEM analysis are included in Fig. 7b, showing a good
explained by considering the stronger link between the agreement between the predicted and experimental results.
outer and inner frames, as observed in Fig. 7b (steeper Fig. 7c includes curves plotted against the load of the sepa-
line). This fact affects the rotation of the mud panel inside rated components of the global deformation. These curves
the frame, with higher impact as that on type-N3, observed correspond to the shear deformation of the mud panels
in Fig. 7c. The difference in the absolute rotation of the (hmud panel obtained from DSP), the absolute rotation of
mud panel, hmud panel, and the global rotation allows larger the mud panels (hmud panel, from DSP), the relative rotation
compression strains on the corners of the mud panels lead- between inner and outer frames (hpanel), and the global shear
ing to higher damage on the panels. This contributes to the deformation curve (hglobal) from the readings of the displace-
decreasing of maximum load and global stiffness. The ment transducers. Each component of the PMWU deforms
absolute stiffness obtained through DSP of the mud panels at different rates. We determined that 11% of the global
does not include the contribution of the frames. deformation is concentrated on the shear deformation of
 B.I. Hassel et al. / International Journal of Sustainable Built Environment 2 (2013) 1–9 7

Fig. 7. Comparison between shear deformation and rotation of the different components of the PMWU. (a) Approximate load-absolute shear
deformation curves, obtained by DSP measured on the surface of the mud panels. (b) Average relative rotation between inner and outer frames obtained
from displacement transducers readings, including the corresponding elastic rotations from FE analysis. (c) Breakdown of the components of deformation
of the PMWU for a representative specimen, type-N3. FEM, finite element method.

the mud panel, (hmud panel), 82% on the absolute rotation of in the shear strain variation, which indicates discontinuity
the mud panels, (hmud panel) and 7% on the relative rotation between these elements.
between frames (hpanel). We conclude that the highest influ- In Fig. 10 we compare the full-field shear strains from
ence on the deformation mechanism of the PMWU is caused DSP and FEM for type-N3, with average strain values of
by the rotation of the mud panel. 1.30  10 3 and 1.55  10 3 respectively, corresponding
Moreover, based on shear strain variation curves from to a load level of 20 kN below the yield point, which is
DSP along the vertical coordinates of the specimens, differ- close to the crack load. Comparison of experimental results
ent load-transmitting behaviors were assumed according to to those predicted by the FE model confirms the correct-
the number of screws, i.e., a difference between types N0, ness of the model.
N1, N2 and types N3, N4 (Fig. 8). In the first group, each In terms of mechanical properties and according to the
has an inner frame and the mud materials behave indepen- bilinear method analysis, linking the mud panels to
dently up to the yield limit of the PMWU. This behavior is the outer frame using three screws (type-N3) produced
different for the second group (e.g., type-N3, Fig. 8). the most efficient PMWU.
The number of connectors highly affected the rotation,
but did not influence the initial shear stiffness of the mud
3.3. Full strain field analysis from FEM and DSP panels. In addition, the mud panel makes the largest contri-
bution to the rotational behavior of the PMWU.
Fig. 9 shows curves of the shear strain variation along a
horizontal line traced at the middle of the sample derived 4. Conclusions
using the FEM and DSP methods at 20 kN. DSP data
correspond to the masking of the mud panel, without The experimental global load-shear deformation (Fig. 5)
homogenization. This is appreciated by the ragged DSP and the analytical results obtained through the FE model
curve, which shows similar tendency to that of the FEM (Fig. 7b), as well as those experimental results compared
prediction. The load transference through the different with the DSP results (Fig. 7c) show consistent agreement
parts of the sample (outer and inner frames) is reflected with the results obtained from the finite element model.
8 B.I. Hassel et al. / International Journal of Sustainable Built Environment 2 (2013) 1–9

Fig. 8. Photographs of test specimens type-N0 and type-N3 superimposed by a shear strain curve along a middle center line. Based on these curves,
obtained from DSP analysis, a different load transmitting behavior related to the number of connectors was assumed.

0.001 analysis as a non-contact method to control the state of

protected ancient buildings. The use of full-field strain
0.0005 Type-3-FEM
Type-N3-FEM
measurements in combination with conventional tech-
Type-3-DSP
Type-N3-DSP
Shear strain

0 niques has proven to increase the accuracy of experimental

MUD
results. The technique allowed us to gain a better under-
-0.0005 standing of the structural behavior of the shear panels,
OUTER FRAME

OUTER FRAME
INNER FRAME

INNER FRAME

by distinguishing the deformation of each constituent part

-0.001
of the PMWU as the rigid-body motion effect was com-
-0.0015
pletely dismissed. Moreover, DSP was especially useful in
0 200 400 600 order to corroborate the FE model.
Horizontal distance along the sample [mm] Each component, frames and panel, was deformed at
Fig. 9. Shear strain variation along a central horizontal line traced along
different rates. The highest influence on the deformation
specimen type-N3 allows comparison of the experimental (DSP) and mechanism of the PMWU was caused by the rotation of
predicted (FEM) results. the mud panel.
Although the number of steel connectors had a great
impact on the rotation of the panels, their influence in
the shear stiffness can be ignored. The mud panel makes
the largest contribution to the rotational behavior of the
PMWU.
The load level that caused the first apparent crack on the
surface of the mud panel was estimated using DSP, which
we denote as LoadF. This load was lower than the yield
strength and about half the strength of the panels, allowing
the PMWU to withstand loads twice as high as the mud
material, without influencing the initial stiffness. This infor-
Fig. 10. Full field shear strain on PMWU when undergoing elastic shear mation is of great value for further development of a
deformation. Specimen type-N3 withstanding a load of 20kN. (a) DSP
method to control the state of existing buildings.
results for a masked area over the mud material. (b) Results predicted by
the FE model, displaying the shear strain of the PMWU. The FEM results Based on the results, further improvement on the FE
agree with our experimental results. model is needed. Some of the parameters considered of
importance are the fiber distributions, the friction coeffi-
Moreover, the concordance shown in Table 1 for the exper- cient between bamboo laths and mud, compression
imental and analytical initial stiffness enforces the later strength of the mud material.
conclusion. Furthermore, the strain field generated by the It is of interest to study the effects of the variation of the
application of shear forces observed through the analytical density of the bamboo mesh located inside the mud panels.
FE model and the DSP measurements are analogous. Namely, the separation between bamboo laths as well as
The combination of experimental and analytical meth- their width.
ods provided information to recommend for practical uses Studying the behavior of the panels under cyclic loading
of the specimen type-N3 as the most efficient PMWU in is mandatory when intending to use them in seismic areas.
terms of the number of connectors. Accelerating weathering tests are of high interest, con-
The combination and comparison of experimental and sidering the low thickness of the panels in comparison with
analytical results show the feasibility of using digital image the traditional walls.
 B.I. Hassel et al. / International Journal of Sustainable Built Environment 2 (2013) 1–9 9

We consider the following as very important applica- Hassel, I., Berard, P., Komatsu, K., 2008. Development of wooden block
tions of the pre-fabricated mud wall units: shear wall – improvement of stiffness by utilizing elements of densified
wood. Holzforschung 62, 584–590.
Hassel, I., Berard, P., Modén, C.S., Berglund, L.A., 2009. The single
- After further study, it is expected to have a system apparatus for shear testing – Full-field strain data and finite element
that allows repairment of shear walls in dwellings analysis of wood in transverse shear. Composite Sci. Technol. 69, 877–
subjected to earthquakes. As the repairment and 882.
even replacement of individual units are possible, HOWTEC, Japan Housing and Wood Technology Center. Tokyo, Japan
Housing and Wood Technology Center. 2004. (In Japanese).
preventing from having to replace the whole struc- Keunecke, D., Sonderegger, W., Pereteanu, K., Luthi, T., Niemz, P., 2007.
ture is made possible. Determination of Young’s and shear moduli of common yew and
- Help maintaining ancient Japanese building tradi- Norway spruce by means of ultrasonic waves. Wood Sci. Technol. 41,
tions alive, by introducing a system that allows 309–327.
flexibility in the construction, control on the Komatsu, K., Kataoka, Y., Mori, T., 2008. Concept of proposed house
and outline for structural performance. Part 1. AIJ J. Technol. 14 (28),
mechanical properties, and no need of highly spe- 447–452 (In Japanese).
cialized craftsmanship. Laroque P, 2007. Design of a low cost bamboo footbridge. Thesis, Master
- Promote the construction with eco-materials and of Engineering in Civil and Environmental Engineering. Massachusetts
energy-saving systems. Institute of Technology.
- Develop and adapt this construction system for MLIT. Housing Bureau, Ministry of Land, Infrastructure and Transport
(2003). The amended Building Standard Law on Sick House Issues. (In
developing countries, utilizing the readily available Japanese). Available at: <http://www.mlit.go.jp/jutakukentiku/build/
raw materials. sickhouse.html>. (Last accessed 06 7, 2012).
MSC Software Vol. A, 2005. Marc Mentat. Volume A: Theory and User
Information, Chapter 8.
Acknowledgements MSC Software Vol. B, 2005. Marc Mentat. Volume B. Finite Elements
Software.
The financial support granted by JSPS (Japan Society MSC Software Vol. C, 2005. Marc Mentat. Volume C: Program Input,
Chapter 3.
for the Promotion of Science) is highly appreciated.
Nakao, M., Yamazaki, Y., 2006. An experimental study on mud-
The authors express their deep appreciation to Professor plastered walls for evaluating ultimate strength and ductility. Part
Lars Berglund, from the Royal Institute of Technology 8 Static incremental analysis by RBSM. Arch. Inst. Jpn. 23 (1),
(KTH), Stockholm, Sweden, for allowing using the digital 241–242.
speckle photography (DSP) equipment and software, Saijo, Y., Kishi, R., Sata, F., Katakura, Y., Urashima, Y., Hatakeyama,
A., Kobayashi, S., Jin, K., Kurahashi, N., Kondo, T., Gong, Y.Y.,
GOM Aramis.
Umemura, T., 2004. Symptoms in relation to chemicals and dampness
We highly appreciate the hard work done by Dr. Ray- in newly built dwellings. Int. Arch. Occup. Environ. Health 77, 461–
mond Wan to proofreading the manuscript. 470.
Shimase, H., Morisako, K., Sato, H., 2007. Hysteresis characteristics of
References mud wall built by traditional plastering in Kyoto. Proc. Annu.
Meeting Kinki Branch AIJ (Architectural Institute of Japan) 47, 333–
336 (In Japanese).
Architecture Institute of Japan, 2002. Standard for Structural design of
Tabuchi, A., Kitamori, A., Mori, T., Komatsu, K., 2006. Lateral shear
timber structure, Tokyo, 104–106.
performance of earth-wall in case of typical town house in Kyoto. J.
Berard, P., Yang, P., Yamauchi, H., Umemura, K., Kawai, S., 2006.
Struct. Construct. Eng. 605, 143–150 (In Japanese).
Modeling of cylindrical LVL: Finite elements models of a flat
Yokobayashi S., Sato M., 2007. A Study on the Heat and Moisture
interlocked LVL and a comparison of standard and nonstandard
Properties of “Tsuchikabe” – A Specialty of Hyogo Prefecture.
testing methods in the elastic domain. In: Dai, C. (Ed.), Proceedings of
Memoirs of the Osaka Institute of Technology. Series A; 52(1),
the 2nd international symposium on veneer processing and products.
1–4.
Vancouver, British Columbia, Canada, pp. 428–435.