Download as pdf or txt
Download as pdf or txt
You are on page 1of 9

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

net/publication/326674720

Joinery connections in timber frames: analytical and experimental explorations


of structural behavior

Conference Paper · July 2018

CITATIONS READS

7 2,123

2 authors:

Demi Fang Caitlin T. Mueller


Massachusetts Institute of Technology Massachusetts Institute of Technology
9 PUBLICATIONS   23 CITATIONS    76 PUBLICATIONS   680 CITATIONS   

SEE PROFILE SEE PROFILE

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

Mechanics, Assembly, and Optimization of Interlocking Timber Joints View project

Geometría, Estructura, Arquitectura View project

All content following this page was uploaded by Demi Fang on 29 July 2018.

The user has requested enhancement of the downloaded file.


Proceedings of the IASS Symposium 2018
Creativity in Structural Design
July 16-20, 2018, MIT, Boston, USA
Caitlin Mueller, Sigrid Adriaenssens (eds.)

Joinery connections in timber frames: analytical and experimental


explorations of structural behavior
Demi FANG*, Caitlin MUELLER*

*Building Technology, Massachusetts Institute of Technology


77 Massachusetts Ave, MIT Room 5-418
Cambridge, MA 02139, USA
dfang@mit.edu

Abstract
Innovations in mass timber have ushered in a resurgence of timber construction. Historic timber structures
feature joinery connections which geometrically interlock, rarely featuring in modern construction which
utilizes steel fasteners for connection details. Research in the geometric potential and mechanical
performance of joinery connections remain disparate. This study seeks to develop a performance-driven
design framework for the geometry of joinery connections. Experimental and analytical models for three
types of joinery connections are presented and compared. The T* type joint, which uses a T-shaped tenon
instead of a dovetail, experimentally showed the highest rotational stiffness. The analytically predicted
rotational stiffness of the T* type joint comes within 20% of the experimentally determined value. A
preliminary parametric study through the analytical model demonstrates how geometric parameters can be
varied to achieve desired rotational stiffness.
Keywords: timber, timber joints, interlocking, connection design, mechanics, rotational stiffness, semi-rigidity

1. Introduction
Reductions in embodied energy made possible by timber construction, which has gained resurgence in the
past two decades, are increasingly valuable for the built environment as urban densities increase and as
building technologies advance towards net-zero operational energy. The design of connections in timber
structures is crucial to global structural performance. Historic wood construction in different parts of the
globe widely used interlocking joints before metal connections (Foliente [1]). In modern construction, two
main contexts in which wood construction might exhibit interlocking connections are 1) in low-rise
structures built in the timber-frame tradition, or 2) in unique structures (e.g. Yusuhara Bridge Museum,
2010, Kengo Kuma and Associates; Tamedia Office Building, 2013, Shigeru Ban Architects). Timber-
frame specialists today typically build such joints at the low-rise residential scale, relying on a combination
of traditional craft and modern specialized analysis. The analysis procedures for these unique wood-only
joints are less standardized than those of metal connectors. The potential benefits of using joinery
connections in mid-rise scale buildings are multifaceted. The elimination of metal joints could further drive
down the embodied energy of structures. Using joinery connections may offer new implications in
construction assembly speed of timber structures. Innovations in digital fabrication suggest that the cost of
milling joinery could become competitive with that of fasteners.

2. Literature review and problem statement


Aside from the Tamedia Office Building, the relationship between the form and performance of joinery
connections remains unexplored in a majority of modern timber construction. A similar gap exists in
research. Interlocking joints as a design concept has piqued interest in a variety of branches of architectural

Copyright © 2018 by Demi FANG, Caitlin MUELLER


Published by the International Association for Shell and Spatial Structures (IASS) with permission.
Proceedings of the IASS Symposium 2018
Creativity in Structural Design

research, including in assembly and in digital fabrication (Tai [2], Page [3], Böhme [4]). In parallel, several
works examine the structural mechanics of interlocking joints, using combinations of analytical and
experimental methods (Brungraber [5], Lang and Fodor [6], Guan et al. [7], Shope [8], Roche [9]). Recent
works published in English have applied Inayama’s embedment theory [10] to analyze rotational stiffness
in joinery connections (Yeo et al. [11], Ogawa et al. [12]).
The work presented in this paper investigates the design of performance-driven geometries for interlocking
joints and assesses their feasibility for mid-rise structures. This work targets the gap between the
architectural research, which is rich in demonstrating the conceptual design potential of interlocking joints;
and the analytical research, which typically focuses on assessing traditional geometries. Specifically, this
paper assesses the rotational stiffness of three types of beam-column joinery connections. The joint behavior
is considered semi-rigid, or an intermediate between an idealized moment connection and an idealized pin
connection. Preliminary joint designs are fabricated and tested to determine the experimental rotational
stiffness of each joint type. An assessment of component behavior within one of the joint types is presented
to develop an analytical model for the joint’s rotational stiffness. Finally, preliminary parametric studies on
the effect of geometry on the joint’s rotational stiffness are presented.

Side view of joint Top view; beam Side view of joint Top view; beam Side view of joint Top view; beam
x1 x2 x3 x1 x2 x3 x1 x2 x3 x1 x2 x3 x1 x2+x3 x1 x2+x3
y1 y1 y1
y2 y2 y2
h y1 h y1 h y1
beam beam beam

post cantilevers post cantilevers post cantilevers

y1 y1 y1
y2 y2 y2
h y1 h y1 h y1

Top view; post Top view; post Top view; post

post post post

(a) T type joint (b) T* type joint (c) D type joint

Figure 1: Three types of joints studied in this work.

3. Experimental testing of rotational stiffness in T, T*, and D type joints


In this test, three joint types were tested: types T, T*, and D (Figure 1). Geometric parameters of the tested
specimens (about 1/3 scale of sections that might appear in mid-rise structures) are summarized in Table 1.

Table 1: Geometric parameters for T and T* type joints

x1 x2 x3 y1 y2 h
1.27 cm (½”) 1.9 cm (¾”) 2.54 cm (1”) 1.27 cm (½”) 1.27 cm (½”) 5.71 cm (2¼”)

3.1 Fabrication and load testing setup


Joints were made by hand using Nordic Lam elements and various woodworking tools (table saw, mortiser, mallet,
chisel) (
Figure 2a). Each joint specimen consisted of a cantilevering beam and a post, with the tenon on the beam and the
mortise at the top of the post. To test the rotational stiffness of the joint, the joint was tested with the post upright
and a vertical load applied to the free end of the beam (
Figure 2b).
The load-displacement testing was done on an MTS test machine with a 50-kN load cell. A total of 6
specimens of type T, 5 specimens of type T*, and 3 specimens of type D were tested.
Material properties used throughout are as follows:
• Stiffness along grain: E0 = 11,030 N/mm2 (1.6106 psi) (Nordic Lam catalog)

2
Proceedings of the IASS Symposium 2018
Creativity in Structural Design

• Stiffness perpendicular to grain: E90 = E0/30 = 365 N/mm2 (5.3104 psi) (Forest Products
Laboratory [13])
• Coefficient of friction:  = 0.5 to 0.7 for wood of intermediate moisture content (Forest Products
Laboratory [13]). In this study  = 0.5 is assumed.
The distance of applied load from the joint (ri) was determined by considering the typical moment-shear
ratio that might be experienced at the joint of a beam with a similar cross-sectional aspect ratio at the mid-
rise scale. For each specimen, raw load-displacement data (P-r) was converted to moment-displacement
(M-) data and the slope of the elastic region of the latter was recorded as the joint’s rotational stiffness.

(a) (b)
Figure 2: (a) Beam-column joint specimens for experimental testing. (b) Experimental setup.

4. Analytical model of rotational stiffness in T* type joint


From experimental results, the most promising of these joint types proved experimentally (see Section 5)
to be the T* joint (Figure 1b). The analytical model for T* type joint is presented in this section. Four types
of rotational springs were identified in the joint: 1) springback from embedment due to angular
displacement, 2) bending deflection from cantilevers on the post, 3) shear deflection from cantilevers on
the post, and 4) differential torsion in beam tenon.
x3 r P Stresses and forces induced on a single cantilever of T* type joint
i
Side view of joint
σ0 σembed
h/6 C0 1/3 (h - xp)
h/2 C1
h - xp
h 2h/3 2h/3
beam h/2 C0 xp σembed
h/6 1/3 xp C2
x=0 σ0 Cs
(3/2)x3 σembed, s

post

compressive stress resultant force compressive stress resultant force


Top view; post
post cantilevers (a) before angular displacement (b) after angular displacement

Figure 3: Stresses and resultant forces induced in a single cantilever of T* type joint due to applied point load.

4.1 Springback from embedment due to angular displacement


This spring considers the embedment stresses that occur after angular displacement occur (Figure 3b).
According to Inayama’s embedment theory and assumptions applied by Kitamori et al. [14],
𝐸90
𝜎𝑒𝑚𝑏𝑒𝑑 (𝑥) = 𝜃𝑥 (1)
𝑍0
embed,s is embedment stress. Z0 is the thickness dimension of the embedded member, which in this case is
x3.  represents the embedment angle and in this case is equivalent to the angular displacement of the
joint, and x represents the coordinate along the x-axis as indicated in Figure 3.

3
Proceedings of the IASS Symposium 2018
Creativity in Structural Design

xp represents the distance from the bottom edge of the beam (x=0) to the neutral axis of embedment
stresses and is solved for using equilibrium. The embedment stress for springback (embed,s(x), Equation 2)
replaces x of embedment stress (Equation 1) with the deformed function 𝛿𝑥 (𝑥) = (𝜃𝑥𝑝 )𝑒 −𝑎𝑥 :
𝐸90 𝐸90
𝜎𝑒𝑚𝑏𝑒𝑑,𝑠 (𝑥) = 𝛿𝑥 (𝑥) = (𝜃𝑥𝑝 )𝑒 −𝑎𝑥 (2)
𝑍0 𝑍0
3 3
Inayama’s embedment theory states that 𝑎 = = 2𝑥 and that embed,s(x) is to be integrated over the
2𝑍0 3
3 3
distance 𝑍 = 𝑥 to determine its resultant force Cs. The resultant forces are then:
2 0 2 3
ℎ−𝑥𝑝
𝑦1 𝐸90 2
𝐶1 (𝜃) = 𝑦1 ∫ 𝜎𝑒𝑚𝑏𝑒𝑑 (𝑥) 𝑑𝑥 = (ℎ − 𝑥𝑝 ) 𝜃 (3)
0 2𝑥3
𝑥𝑝
𝑦1 𝐸90 2
𝐶2 (𝜃) = 𝑦1 ∫ 𝜎𝑒𝑚𝑏𝑒𝑑 (𝑥) 𝑑𝑥 = (𝑥𝑝 ) 𝜃 (4)
0 2𝑥3
3
2 𝑥3 2 9
𝐶𝑠 (𝜃) = 𝑦1 ∫ 𝜎𝑒𝑚𝑏𝑒𝑑,𝑠 (𝑥) 𝑑𝑥 = (1 − 𝑒 −4 ) 𝑦1 𝐸90 𝑥𝑝 𝜃 (5)
0 3
Substituting Equations 3, 4, and 5 into the equilibrium equation 𝐶1 = 𝐶2 + 𝐶𝑠 gives xp (Equation 6). The
value is substituted back into Equations 3, 4, and 5 to determine the embedment forces.
3ℎ
𝑥𝑝 = 9
4𝑥3 (6)
(1 − 𝑒 −4 ) + 6

Friction forces, each magnitude F, arise from embedment forces in the perpendicular direction (Equation
7). The neutral axis in this perpendicular direction is assumed from experimental observation.
𝐹 = 𝜇(𝐶2 + 𝐶𝑠 ) (7)
Finally, the relationship between moment M and angular displacement  may be obtained:
2 2
𝑀 = 𝐶1 × (ℎ − 𝑥𝑝 ) + 𝐶2 × 𝑥𝑝 + 𝐶𝑠 × 𝑥𝑝 + 𝐹 × 𝑦1 (8)
3 3
Substituting Equations 3, 4, 5, 6, and 7 into the above expression gives 𝑀(𝜃) = 𝑘1 𝜃, where k1 represents
the rotational stiffness from embedment springback at one post cantilever.

4.2 Bending deflection from cantilevers on the post


This spring considers the embedment stresses that occur before angular displacement occur (Figure 3a).
These stresses and forces are purely a function of Pri, where P is the applied downward vertical force and
ri is its distance from the neutral axis of the joint (determined from experimental observation). The stresses
and forces can be written as follows, where Pri is assumed to be split evenly into two forces of magnitude
C0 acting on each of the two cantilevers:
3
𝐶0 = 𝑃𝑟 (9)
4ℎ 𝑖
3
𝜎0 = 2 𝑃𝑟𝑖 (10)
ℎ 𝑦1
The rotational stiffness of each cantilever as a result of bending deflection under the applied compression
stresses can now be determined. An x-axis is set to run along the length of the cantilever starting from its
base. The applied load q(x) from compressive stress can be written as a linear function of x, and subsequent

4
Proceedings of the IASS Symposium 2018
Creativity in Structural Design

integrations with respect to x give shear V(x) and moment M(x) distributions, with initial values of V and M
determined via equilibrium (V(0)=0, M(0)=C0h/2). These distributions are shown in Figure 4.

Figure 4: Loading, shear, and moment diagram for one post cantilever of the T* type joint.

𝑑2 𝑤 𝑑𝜃
Assuming pure bending, the relationship 𝑀(𝑥) = −𝐸0 𝐼 𝑑𝑥 2 = −𝐸0 𝐼 𝑑𝑥𝑐 can then be utilized to find the
joint’s rotational stiffness, where c(x) is the angular displacement of the cantilever along its length. To
relate this angular displacement c(x) to the joint’s angular displacement , it is assumed that c(x=h/2)
gives . Note that the cantilever support boundary condition gives (0)=0.
1 1 1 4 1 3 𝐶0 ℎ
𝜃𝑐 (𝑥) = − ∫ 𝑀(𝑥)𝑑𝑥 = − [𝜎0 𝑦1 (− 𝑥 + 𝑥 )+ 𝑥] (11)
𝐸0 𝐼 𝐸0 𝐼 12ℎ 6 2
ℎ 1 ℎ3 ℎ3 𝐶0 ℎ2 𝑃𝑟𝑖 15ℎ
𝜃𝑐 ( ) = − [𝜎0 𝑦1 (− + )+ ]=− ( ) (12)
2 𝐸0 𝐼 192 48 4 𝐸0 𝐼 64
1
Substituting 𝐼 = 12 𝑦1 (𝑥3 )3 , Pri=M, and c(h/2)=  in the above expression, followed by some rearranging,
gives an expression of the form 𝑀 = 𝑘2 𝜃. This stiffness k2 is for one of the two cantilevers.

4.3 Shear deflection from cantilevers on post


The rotational stiffness of each cantilever as a result of shear deflection v(x) can be determined from the
𝑑𝑣 𝛼𝑉(𝑥)
shear distribution V(x) across the cantilever (Figure 4) using the expression 𝑑𝑥 = 𝜃𝑐 (𝑥) = 𝐺𝐴 , where  =
3/2 for rectangular cross-sections. It is assumed that c(x=h/2) gives the joint’s angular deflection .
𝛼𝑉(𝑥) 3 1
𝜃𝑐 (𝑥) = = 𝜎0 𝑦1 (− 𝑥 2 + 𝑥) (13)
𝐺𝐴 2𝐺𝐴 2ℎ
ℎ 3 ℎ ℎ 9ℎ 27
𝜃𝑐 ( ) = 𝜎 𝑦 (− + ) = 𝜎 𝑦 = 𝑃𝑟 (14)
2 2𝐺𝐴 0 1 8 2 16𝐺𝐴 0 1 16𝐺𝐴ℎ 𝑖
Substituting 𝐴 = 𝑦1 𝑥3 , Pri=M, and c(h/2)=  in the above expression, followed by some rearranging, gives
an expression of the form 𝑀 = 𝑘3 𝜃. This stiffness k3 is for one of the two cantilevers.

4.4 Differential torsion in beam tenon


Each flange of the “T”-shape on the beam tenon acts as a small beam of length y1 under torsion from the
compressive reaction supplied by the post’s cantilever (Figure 5). The angle of twist θt of a beam with a
𝑑𝜃
non-circular section can be determined by 𝑇 = 𝐺𝐽 𝑑𝑥𝑡 , with T as torsion experienced by the beam and J a
geometric section property. The compressive reaction C0 acts at x = y1/2; the angle of twist due to C0 is thus
𝑦1 𝑇 𝑦1
𝜃𝑡 ( ) = × (15)
2 𝐺𝐽 2

5
Proceedings of the IASS Symposium 2018
Creativity in Structural Design

For a solid rectangular section with side lengths a and b, where a > b, section property J is given by =
1 𝑏 𝑏4
𝑎𝑏3 (3 − 0.21 𝑎 (1 − 12𝑎4)) , where a represents the longer section dimension. Substituting a = h, b = x2,
ℎ 3 ℎ 𝑃𝑟
𝑇 = 𝐶0 3 = 4ℎ 𝑃𝑟𝑖 3 = 4 𝑖, Pri=M, and t =  into Equation 15, followed by some rearranging, gives an
expression of the form 𝑀 = 𝑘4 𝜃. This stiffness k4 is for one of the two cantilevers.
x3 r P compressive stress compressive stress resultant force in Top view; beam
i
in post in beam flange beam flange
σ0 σ0
h/2 C0 h/3
h y1 /2
h/2 C0
beam
σ0

post Side view of joint

Figure 5: Torsion experienced by flanges of beam tenon.

4.5 Assembly of spring stiffnesses for rotational stiffness of T* type joint


Each of the four spring stiffnesses calculated above represents the rotational stiffness for one of the two
cantilevers on the post. For each spring, the effect of stiffness of both cantilevers can be determined by
considering the springs to be acting in parallel. On the other hand, the 4 types of springs act in series. The
final rotational stiffness k of the T* type joint can be calculated as follows:
1
𝑘=
1 1 1 1 (16)
+ + +
2𝑘1 2𝑘2 2𝑘3 2𝑘4

4.6 Preliminary parametric study

Variation of Joint Rotational Stiffness with x 2


stiffness (N-m/rad)
Joint rotational

5000

0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
x2/(x2+x3)

Figure 6: Relationship between geometric parameter x2 and analytical prediction of joint stiffness in T* type joint

In the presented analytical model, the relationship between each geometric parameter and the joint
rotational stiffness is not immediately apparent due to the bulk of intermediate calculations. Parametric
tools such as Microsoft Excel or Grasshopper enable preliminary sensitivity studies. For example, the effect
of x2 on k was explored by fixing the distance x2+x3 and varying x2. According to the model, joint stiffness
can be maximized when the proportion of x2 over x2+x3 is 0.34 (Figure 6). More generally, the plotted
relationship demonstrates how geometric parameters can be varied to achieve varying rotational stiffnesses.

5. Results
The experimental data for the elastic regions of joint types T, T*, and D are plotted in Figure 7. The
analytically predicted rotational stiffness of joint type T* for specimens tested experimentally (based on
geometric parameters given in Table 1) is given in Table 2.

6
Proceedings of the IASS Symposium 2018
Creativity in Structural Design

Moment-Displacement Behavior of Joints


250

200

Moment (N-m)
150

100

50

0
0 0.02 0.04 0.06 0.08 0.1 0.12
Angular Displacement (rad)
T T* D

Figure 7: Moment-displacement curves of elastic regions of specimens resulting from experimental tests.

Table 2: Analytical determination of rotational stiffness in T*

Spring k1 k2 k3 k4 Joint rotational stiffness k


Rotational stiffness (N-m/rad) 6050 9740 5300 42,400 4160

Figure 8 plots the calculated rotational stiffness of each experimental specimen. The T* type joint
demonstrated the highest rotational stiffness of the three types. Perhaps unexpectedly – due to its
widespread use in furniture making and carpentry – the dovetail geometry of the D type joint exhibited the
lowest rotational stiffness with its mean (96 N-m/rad) only equivalent to 2.8% of that of the T* type joint.
Qualitatively, it was observed that the ductile behavior of the D type joint was a result of outward spreading
of its post cantilevers, a behavior not exhibited by the T or T* type joints. While it might be intuited that
the T type joint would provide a higher rotational stiffness over the T* type joint due to the extra confining
tab of the T type joint, the experimental results demonstrate the opposite (35% higher rotational stiffness
in the T* type joint). Further work in developing an analytical model for the T type joint in comparison to
the one presented for the T* type joint is required to explain this experimental result.
The analytically predicted rotational stiffness for the T* type joint is also indicated in Figure 8. The
analytical prediction (4160 N-m/rad) is 20% higher than the mean experimentally determined stiffness.
2560 N-m/rad
T 3460 N-m/rad 4160 N-m/rad
T*
96.0 N-m/rad
D
0 500 1000 1500 2000 2500 3000 3500 4000 4500
Rotational stiffness (N-m/rad)

experimental result experimental mean analytical prediction

Figure 8: Experimentally determined rotational stiffnesses

6. Conclusion and further work


The experimental results of loading joinery connections (Figure 7) showed that the D type joint (dovetail
tenon) gave the lowest rotational stiffness in comparison to the T and T* type joints (T-shaped tenons). The
T* type joint exhibited a higher rotational stiffness despite lacking the confining tab of the T type joint. An
analytical model for the T* type joint was developed by identifying and calculating spring stiffnesses in the
joint. The analytically predicted rotational stiffness of the T* type joint (Table 2) was 20% higher than its
mean experimental rotational stiffness (Figure 8); the predicted model is close but may require further work
to refine the spring stiffness analysis. The analytical model provides a viable framework for a parametric
study between geometry and joint rotational stiffness (Figure 6). Because deflections often control member

7
Proceedings of the IASS Symposium 2018
Creativity in Structural Design

design in timber structures, higher rotational stiffnesses can be desirable for material efficiency in multi-
story structures; it is valuable to predict how geometric variations increase joint rotational stiffness.
Future experimental tests will use digitally fabricated joints to not only reduce variation among specimens
but also generate larger sample sizes. Further assessment of joint strength is also of interest.
Exploring the relationship between geometry and performance for joinery connections provides a step
towards the design of interlocking joints for multi-story structural frames. The beam of the joint prototypes
in the present work could be envisioned as the primary load-bearing beam on which secondary beams rest,
transferring load into columns. Future work could examine the effect of stacking frames or the potential of
the joint to receive a perpendicular beam, considering global applications in multi-story structural frames.

Acknowledgements
The authors thank Tedd Benson & Andrew Dey (Bensonwood), Julie Frappier (Nordic Structures),
Christopher Dewart & Steve Rudolph (MIT), Yijiang Huang, Felix Amtsberg, Paul Mayencourt.

References
[1] G. C. Foliente, “History of Timber Construction,” ASTM Int., Jan. 2000.
[2] A. S.-C. Tai, “Design for assembly : a computational approach to construct interlocking wooden
frames,” Thesis, Massachusetts Institute of Technology, 2012.
[3] M. Page, “A Robotic Fabrication Methodology for Dovetail and Finger Jointing: An Accessible &
Bespoke Digital Fabrication Process for Robotically-Milled Dovetail & Finger Joints,” in
Proceedings of the 37th Annual Conference of the Association for Computer Aided Design in
Architecture (ACADIA), Cambridge, MA, 2017, pp. 456–463.
[4] L. F. G. Böhme, F. Q. Zapata, and S. M. Ansaldo, “Roboticus tignarius: robotic reproduction of
traditional timber joints for the reconstruction of the architectural heritage of Valparaíso,” Constr.
Robot., pp. 1–8, Jul. 2017.
[5] R. L. Brungraber, “Traditional timber joinery : a modern analysis,” Stanford University, 1985.
[6] E. M. Lang and T. Fodor, “Finite Element Analysis of Cross-halved Joints for Structural Composites,”
Wood Fiber Sci., vol. 34, no. 2, pp. 251–265, Jun. 2007.
[7] Z. W. Guan, A. Kitamori, and K. Komatsu, “Experimental study and finite element modelling of
Japanese ‘Nuki’ joints — Part two: Racking resistance subjected to different wedge configurations,”
Eng. Struct., vol. 30, no. 7, pp. 2041–2049, Jul. 2008.
[8] M. G. Shope, “Strength characterization of wood to wood connections using stress field analysis,”
Thesis, Massachusetts Institute of Technology, 2016.
[9] S. N. Roche, “Semi-Rigid Moment-Resisting Behavior of Multiple Tab-and-Slot Joint for Freeform
Timber Plate Structures,” PhD, Ecole Polytechnique Federale de Lausanne, Lausanne, Switzerland,
2017.
[10] M. Inayama, “Wooden embedment theory and its application,” PhD, University of Tokyo, Tokyo,
Japan, 1991.
[11] S.-Y. Yeo, K. Komatsu, M.-F. Hsu, and Z. Que, “Mechanical model for complex brackets system of
the Taiwanese traditional Dieh-Dou timber structures,” Adv. Struct. Eng., vol. 19, no. 1, pp. 65–85,
Jan. 2016.
[12] K. Ogawa, Y. Sasaki, and M. Yamasaki, “Theoretical modeling and experimental study of Japanese
‘Watari-ago’ joints,” J. Wood Sci., vol. 61, no. 5, pp. 481–491, Oct. 2015.
[13] Forest Products Laboratory, “Wood Handbook: Wood as an Engineering material,” U.S. Department
of Agriculture, Forest Service, Forest Products Laboratory, Madison, WI, General Technical Report
FPL-GTR-190, Apr. 2010.
[14] A. Kitamori, “Embedment theory and Nuki-joint (Section 4.4),” Fundam. Theory Timber Eng. Archit.
Inst. Jpn. AIJ, pp. 97–103.

View publication stats

You might also like