Timber joinery in modern construction:

Mechanical behavior of wood-wood connections

By

Demi L. Fang

Bachelor of Science in Engineering, Civil and Environmental Engineering

Princeton University, 2017

Submitted to the
Department of Architecture
in Partial Fulfillment of the Requirements for the Degree of

Master of Science in Building Technology

at the

Massachusetts Institute of Technology

May 2020

© 2020 Demi Fang. All rights reserved.

The author hereby grants to MIT permission to reproduce and to distribute publicly paper and
electronic copies of this thesis document in whole or in part in any medium now known or
hereafter created.

Signature of Author: __________________________________________________________________

Department of Architecture
08 May 2020
Certified by: _________________________________________________________________________
Caitlin Mueller
Associate Professor, Departments of Architecture and Civil and Environmental Engineering
Thesis Supervisor
Accepted by: ________________________________________________________________________
Leslie K. Norford
Chair, Department Committee on Graduate Students
2
 Timber joinery in modern construction:
Mechanical behavior of wood-wood connections
By

Demi L. Fang

Submitted to the Department of Architecture on May 8, 2020

in Partial Fulfillment of the Requirements for the Degree of
Master of Science in Building Technology

Abstract
Timber joinery is a method of geometrically interlocking timber elements prevalent in historic
cultures around the world, including North America, Europe, and East Asia. The use of joinery as
structural connections faded with the development of metallic screws and nails. Two recent
developments offer the opportunity to revive this historic timber connection type: 1) the
increasing desire to reduce embodied carbon in buildings by replacing more components with
timber as a low-carbon structural material, and 2) recent digital fabrication capabilities which
enable the precise milling of complex geometries as an alternative to the time- and labor-
intensive handiwork required previously.
How can joinery connections be designed in modern structural joints? Can we quantify the
sustainability advantage of using these all-timber joints in lieu of the modern convention of
metallic fasteners? This thesis addresses both questions as applied to the Japanese Nuki joinery
type, though the workflows may be applied to any joinery geometry. First, the rotational stiffness
of the Nuki joint is characterized and cross-verified using multiple methods. Second, the
embodied carbon of a gravity frame using Nuki joints is compared to that of a gravity frame
using conventional metallic fasteners. The use of Nuki joints not only eliminates the use of steel
and aluminum but also provides rotational stiffnesses that enable smaller beam sections to be
used. It is shown that gravity frames designed with Nuki joints could reduce embodied carbon
by over 70% compared to gravity frames designed using conventional beam hanger connections.
The findings make a case for all-timber joinery connections to be implemented as a sustainable
alternative to conventional metallic connections used in modern timber construction.

Thesis Supervisor: Caitlin Mueller

Associate Professor, Departments of Architecture and Civil and Environmental Engineering

3
4
Acknowledgements
I have to begin by expressing my utmost gratitude to my adviser Prof. Caitlin Mueller for her
outstanding mentorship in all these years I have been fortunate to work with her. Thank you for
being endlessly supportive of my professional growth and ambitions and for always trusting in
my potential as a researcher. I very much look forward to continuing to work with you in the next
few years!
Much of this research could not have been possible without my collaborators for the mechanical
analyses of the Nuki joint. A fortuitous encounter at IASS 2018 led to a rich collaboration with
Julieta Moradei, Aliz Fischer, Nick Sherrow-Groves, and Dr. Benshun Shao at Arup San Francisco;
as well as Jan Brütting and Prof. Corentin Fivet at the Structural Xploration Lab of EPFL. I thank
and applaud Danny Landez, our undergraduate research assistant funded by the MIT UROP, for
independently carrying out experiments during my gap year.
I am thankful to Dr. Ben Brungraber for early discussions on the topic of analyzing timber frames,
as well as to Andrew Dey at Unity Homes and Tedd Benson at Bensonwood for their hospitality
in showing us their facilities. Many thanks go to Nordic Structures (in particular Jean-Luc Dubois)
and Bensonwood for coordinating to provide us glulam offcuts for our prototypes. Our
prototypes were fabricated in the MIT N51 wood shop (thank you, Chris Dewart) and the
Autodesk BUILD Space in Boston, and then tested in the Course 1 load-testing labs (thank you,
Steve Rudolph). In Course 1, I additionally thank Prof. Tal Cohen for supervising my research on
the dynamics of joinery connections in historic structures, as well as Prof. Josephine Carstensen
for early discussions on possible applications of level-set topology optimization. I also appreciate
interesting early discussions and insights shared by Dr. Felix Amtsberg.
I primarily want to thank my colleague Dr. Paul Mayencourt for sharing camaraderie, valuable
engineering insights, and an enthusiasm for timber and its craft with me throughout the past
three years. I am also grateful for MIT Architecture’s furniture class with Chris Dewart and joinery
workshops with the Suikoushya school (Kawai-sensei and Dylan Iwakuni, whose visits in early
2020 were sponsored by MISTI Japan) which have introduced me to the beauty of joinery
craftsmanship.
I deeply thank Meredith and Ray Rothrock for their warm support of my endeavors during our
work on the greenhouse in 2018. I am also thankful to my teammates and mentors at SOM
Chicago, who graciously hosted me for a fruitful gap year internship in 2019. The opportunity to
immerse in a vibrant environment of multidisciplinary design in the applied field helped me
incubate several ideas that emerged in this thesis.
During both joyful and difficult times at MIT, but especially during this global pandemic, I am
thankful for my support networks at MIT, in the Cambridge area, and beyond. My most heartfelt
gratitude goes out to my family: to soon-to-be Dr. Billy Fang for always serving as an outstanding
role model to me, and to Mom and Dad for their generous and unwavering support of my growth
and abilities.
Finally, I am very grateful that my research at MIT was made possible by the MIT Presidential
Graduate Fellowship and SA+P’s J. A. Curtis (1953) Fund.

5
6
Table of Contents
Abstract ........................................................................................................................................... 3

Acknowledgements .......................................................................................................................... 5

Publications and presentations related to thesis ............................................................................ 11

List of mathematical symbols ......................................................................................................... 13

List of figures ................................................................................................................................. 15

List of tables................................................................................................................................... 19

1 Introduction ............................................................................................................................ 21
1.1 Motivation ................................................................................................................................... 21
1.2 Timber joinery in history.............................................................................................................. 24
1.3 Timber joinery in modern structures ........................................................................................... 27
1.4 Potential benefits and opportunities .......................................................................................... 30
1.4.1 Non-destructive assembly and Design for Deconstruction ................................................ 30
1.4.2 Seismic systems ................................................................................................................... 31
1.4.3 Reduced embodied carbon ................................................................................................ 31
1.4.4 Biophilia............................................................................................................................... 32
1.4.5 Digital fabrication ................................................................................................................ 32
1.5 Thesis aims .................................................................................................................................. 32

2 Literature review ..................................................................................................................... 33

2.1 Static analysis .............................................................................................................................. 33
2.1.1 Analytical methods .............................................................................................................. 35
2.1.2 Experimental methods ........................................................................................................ 36
2.1.3 Numerical methods ............................................................................................................. 36
2.1.4 Static behavior of mortise-and-tenon-type joinery ............................................................. 36
2.2 Dynamic analysis ......................................................................................................................... 37
2.3 Digital fabrication and joinery ..................................................................................................... 38
2.4 Evaluating sustainability of early-stage structural design exploration using embodied carbon 38
2.5 Problem statement ...................................................................................................................... 39
2.6 Thesis format ............................................................................................................................... 39

3 Methodology for characterizing the static mechanical behavior of the Nuki joint................... 41
3.1 Analytical method: general approach ......................................................................................... 41
3.2 Analytical method: Nuki joint ...................................................................................................... 43
3.2.1 Elastic embedment ............................................................................................................. 45
3.2.2 Plastic embedment ............................................................................................................. 46
7
 3.2.3 Contribution of friction ........................................................................................................ 49
3.2.4 Determining rotational stiffness .......................................................................................... 49
3.3 Experimental methods ................................................................................................................ 50
3.4 Numerical methods ..................................................................................................................... 51
3.5 Statement of contributions ......................................................................................................... 53
3.6 Summary ..................................................................................................................................... 53

4 Static mechanical behavior of the Nuki joint ........................................................................... 55

4.1 Material properties ...................................................................................................................... 56
4.2 Results: Douglas fir prototypes ................................................................................................... 58
4.2.1 Analytical results.................................................................................................................. 58
4.2.2 Experimental results ............................................................................................................ 59
4.2.3 Comparison of results ......................................................................................................... 63
4.3 Results: glulam prototypes.......................................................................................................... 66
4.3.1 Analytical results.................................................................................................................. 66
4.3.2 Experimental results ............................................................................................................ 67
4.3.3 Numerical results ................................................................................................................. 71
4.3.4 Comparison of results ......................................................................................................... 72
4.4 Discussion ................................................................................................................................... 76
4.4.1 Comparison of overall moment-rotational-displacement behavior .................................... 76
4.4.2 Comparison of stiffness values ............................................................................................ 76
4.4.3 Comparison of relative increases in rotational stiffness from increasing beam width ........ 77
4.4.4 Comparison of yield moment values .................................................................................. 77
4.4.5 Limitations in experimental testing ..................................................................................... 77
4.5 Conclusions ................................................................................................................................. 78

5 Nuki joint in modern timber construction ............................................................................... 79

5.1 Modern mass timber connections............................................................................................... 79
5.2 Structural assumptions ................................................................................................................ 80
5.3 Beam design ............................................................................................................................... 81
5.3.1 Steel beam design .............................................................................................................. 84
5.3.2 Beam design with beam hangers........................................................................................ 84
5.3.3 Beam design with Nuki joints.............................................................................................. 87
5.4 Comparing embodied carbon for all gravity frame design methods ......................................... 91
5.4.1 Assumptions ........................................................................................................................ 91
5.4.2 Quantities ............................................................................................................................ 92
5.5 Discussion: comparing quantities across frame designs............................................................. 98
5.6 Conclusions ............................................................................................................................... 100

6 Conclusions ...........................................................................................................................103
6.1 Summary of contributions ......................................................................................................... 103
6.2 Potential impact ........................................................................................................................ 104
6.3 Limitations and future work ....................................................................................................... 105
8
 6.4 Concluding remarks .................................................................................................................. 105

References ....................................................................................................................................107

Appendix 1: Analytical models for the Nuki joint ..........................................................................113

Part 1: Set up parameters and definitions ............................................................................................ 113
Part 2: Calculate moment as a function of rotational displacement ..................................................... 115
Part 3: Assemble M as a function of theta for some range of theta ..................................................... 118

Appendix 2: Scans of tested prototypes .......................................................................................121

Douglas fir prototypes .......................................................................................................................... 121
Glulam prototypes ................................................................................................................................ 128

Appendix 3: Beam with partially rotational restrained supports ....................................................133

Appendix 4: Beam hanger volume estimates ................................................................................139

Ricon S VS 140x60 ................................................................................................................................ 139
Ricon S VS 200x80 ................................................................................................................................ 140
MEGANT 310x100 ................................................................................................................................ 141
Ricon XL 390x80.................................................................................................................................... 143

9
10
Publications and presentations related to thesis
Fang, Demi, and Caitlin Mueller. 2018. “Joinery Connections in Timber Frames: Analytical and
Experimental Explorations of Structural Behavior.” In Proceedings of the International
Association for Shell and Spatial Structures (IASS) Symposium. Cambridge, MA, USA.
Fang, Demi, Caitlin Mueller, Jan Brütting, Corentin Fivet, and Julieta Moradei. 2019. “Rotational
Stiffness in Timber Joinery Connections: Analytical and Experimental Characterizations of the
Nuki Joint.” In Structures and Architecture: Bridging the Gap and Crossing Borders, edited by
Paulo J.S. Cruz, 229–36. Lisbon, Portugal: CRC Press. [Link]
28.
Fang, Demi, Julieta Moradei, Jan Brütting, Aliz Fischer, Daniel K Landez, Benshun Shao, Nick
Sherrow-Groves, Corentin Fivet, and Caitlin Mueller. 2019. “Modern Timber Design Approaches
for Traditional Japanese Architecture: Analytical, Experimental, and Numerical Approaches for
the Nuki Joint.” In Proceedings of the International Association for Shell and Spatial Structures
(IASS) Symposium. Barcelona, Spain.
Fang, Demi, Alpha Arsano, Nathan Brown, Christoph Reinhart, and Caitlin Mueller. 2019.
“Design Space Exploration for High-Performance Greenhouse Design.” In Proceedings of the
International Association for Shell and Spatial Structures (IASS) Symposium. Barcelona, Spain.

11
12
List of mathematical symbols
𝑏𝑏 beam width (Section 5)
𝐵𝐵𝑑𝑑 beam depth (Sections 3 and 4)
𝐵𝐵𝑤𝑤 beam width (Sections 3 and 4)
𝐶𝐶𝑑𝑑 column depth
𝐶𝐶𝑤𝑤 column width
𝑑𝑑 diameter
∆ linear displacement
𝐸𝐸0 elastic modulus of a timber material parallel to grain
𝐸𝐸90 elastic modulus of a timber material perpendicular to grain
𝜀𝜀𝑦𝑦 yield strain for a timber material under compression perpendicular to grain
ℎ beam depth (Section 5)
𝐼𝐼 area moment of inertia
𝑘𝑘 rotational stiffness of joint
𝑙𝑙 various embedment lengths
𝐿𝐿 various lengths (Sections 3 and 4, Appendix 1); beam span (Section 5, Appendix 3)
𝜇𝜇 coefficient of static friction
𝑀𝑀(𝑥𝑥) bending moment along beam’s long axis 𝑥𝑥
𝑀𝑀𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 allowable moment
𝑀𝑀𝑦𝑦 yield moment
𝑁𝑁 normal contact force
𝑃𝑃𝑅𝑅90 stiffness plastic reduction factor for compression perpendicular to grain
𝑞𝑞 applied line load
𝜎𝜎 bending stress
𝜏𝜏 shear stress
𝜃𝜃 angular displacement
𝑉𝑉(𝑥𝑥) shear of beam along beam’s long axis 𝑥𝑥
𝑤𝑤(𝑥𝑥) deflection of beam perpendicular to long axis 𝑥𝑥
𝑍𝑍 depth of material experiencing embedment
𝑍𝑍0 initial depth of material experiencing embedment

13
14
List of figures
Figure 1. Steel connections between glulam elements at the Bullitt Center (Miller Hull Partnership;
Seattle, WA, USA; 2013). Left: during construction (Court 2012), photo by John Stamets; right:
interior after completion (Schuchart n.d.). .............................................................................. 22
Figure 2. Interior of the Promega Feynman Center “The Crossroads” manufacturing facility (Uihlein-
Wilson Architects, Inc.; Madison, WI, USA; 2013). (WoodWorks n.d.) ..................................... 22
Figure 3. Structural connection detail at T3 (Michael Green Architecture and DLR Group; Minneapolis,
MN, USA; 2016). (Brownell 2016) ............................................................................................ 23
Figure 4. Structural detailing of the Brock Commons Tallwood House (Acton Ostry Architects;
Vancouver, BC, Canada; 2017) at University of British Columbia. (naturally:wood 2016) ........ 23
Figure 5. Connection details of zipper trusses at the John W. Olver Design Building (Leers Weinzapfel
Associates; Amherst, MA, USA; 2017) at University of Massachusetts Amherst. (Miller 2017) . 24
Figure 6. Bracket set joinery at Nandaimon, or the Great South Gate (Nara, Japan; 12th century) of
Todai-ji. Photo by Deanna MacDonald. .................................................................................. 25
Figure 7. Joinery types in Chinese traditional timber architecture. (Shiping 1991) ........................ 25
Figure 8. Cruck framing, an English carpentry technique. Leigh Court Barn is the largest cruck framed
structure in Britain (Worcester, England; 1325). Photo by Simon Webb. ................................ 26
Figure 9. A modern example of North American timber framing tradition. Photo by Vermont Timber
Works Inc. ............................................................................................................................... 26
Figure 10. Example joinery from the North American timber framing tradition. (Benson 1997) ..... 27
Figure 11. Joinery detail at the Yusuhara Wooden Bridge Museum (Kengo Kuma & Associates;
Yusuhara, Japan; 2011). Photo by Takumi Ota. ...................................................................... 28
Figure 12. All-timber structural system of the Tamedia Office Building (Shigeru Ban Architects; Zurich,
Switzerland; 2013). Photo by Didier Boy de la Tour. ............................................................... 28
Figure 13. "Cat's paw" timber tension joinery detail at the Writers Theatre (Studio Gang; Glencoe,
IL, USA; 2016). Photo by Steve Hall © Hedrich Blessing. ........................................................ 29
Figure 14. Timber construction companies like Bensonwood create modern homes, like this Vermont
lake house (Ludlow, VT, USA; 2008) in the traditional style of North American timber framing.
Photo by Al Karevy. ................................................................................................................ 30
Figure 15. "The Bed" from Thuma is marketed for its ease of assembly marked by an absence of
metal hardware. (“The Bed” n.d.) ........................................................................................... 31
Figure 16. The traditional Japanese Nuki geometry as a mortise-and-tenon-type joinery connection
for beams and columns. .......................................................................................................... 37
Figure 17. These illustrations show the surface deformation of the indirectly loaded area just outside
of the area under direct contact with embedment. (c) through (e) illustrate the case of rotational
partial compression from different angles. Figure by (Roche 2017), synthesizing images from
(Inayama 1991; 1993; Architectural Institute of Japan (AIJ) 2010). ......................................... 42
Figure 18. Analytical model (as a cut section down the middle of the prototype width): (a) joint
geometry dimensions, (b) elastically compressed volumes, and (c) compressed volumes for an
elastoplastic material behavior. Elastically compressed volumes are shown in blue tones; plastic
volumes are given in magenta tones. The boxed diagrams in (b) and (c) represent the contact

15
 at the ended beam end and thus the subscripts e are used; the diagrams are analogous for the
continuous beam end, where subscripts e would be replaced with subscripts c. Note that the
formula for L has been updated since the original published study. Adapted from original
illustration by collaborator Jan Brütting. ................................................................................. 44
Figure 19. Setup for experimental load-testing of Nuki joint prototypes. Graphic made with
collaborator Jan Brütting. ....................................................................................................... 51
Figure 20. Mortise with rounded corners in the digitally fabricated prototypes. ............................ 51
Figure 21. Progressive frames from testing rotational stiffness of glulam specimens. .................... 51
Figure 22. (a) Experimental setup for Nuki joint testing, and (b) boundary and loading conditions for
numerical simulation. Figure by collaborators Aliz Fischer and Jan Brütting. ......................... 52
Figure 23. Stress in the beam perpendicular to grain (psi) throughout the numerical simulation for
the B w = 1 in. specimen. Analysis performed by Aliz Fischer, Benshun Shao, and Nick Sherrow-
Groves of Arup San Francisco. ................................................................................................ 52
Figure 24. Nuki joint and associated dimensions examined in the case study. .............................. 56
Figure 25. Generalized stress-strain behavior of timber under compression perpendicular to grain,
and associated material values. .............................................................................................. 57
Figure 26. Moment-rotational-displacement curve for the Douglas fir Nuki joint determined by the
analytical model. ..................................................................................................................... 58
Figure 27. Moment-rotational-displacement curves for the Douglas fir Nuki joint with B w of 1”, as
determined by experimental testing. ...................................................................................... 59
Figure 28. Summary of rotational stiffnesses of all prototypes of the Douglas fir Nuki joint with B w of
1”, as determined by experimental testing. ............................................................................ 60
Figure 29. Moment-rotational-displacement curves for the Douglas fir Nuki joint with B w of 1.5”, as
determined by experimental testing. ...................................................................................... 61
Figure 30. Summary of rotational stiffnesses of all prototypes of the Douglas fir Nuki joint with B w of
1.5”, as determined by experimental testing. ......................................................................... 61
Figure 31. Moment-rotational-displacement curves for the Douglas fir Nuki joint for all beam widths,
as determined by experimental testing. ................................................................................. 62
Figure 32. Comparison of moment-rotational-displacement curves using different methods for the
Douglas fir Nuki joint with a B w of 1”. ..................................................................................... 63
Figure 33. Comparison of moment-rotational-displacement curves using different methods for the
Douglas fir Nuki joint with a B w of 1.5”. .................................................................................. 64
Figure 34. Comparison of rotational stiffnesses of the Douglas fir Nuki joint as determined by
different methods, with (a) B w of 1” and (b) B w of 1.5”. FEA was not carried out for these Douglas
fir prototypes. ......................................................................................................................... 65
Figure 35. Moment-rotational-displacement curve for the glulam Nuki joint determined by the
analytical model. ..................................................................................................................... 67
Figure 36. Moment-rotational-displacement curves for the glulam Nuki joint with B w of 1”, as
determined by experimental testing. ...................................................................................... 68
Figure 37. Summary of rotational stiffnesses of all prototypes of the glulam Nuki joint with B w of 1”,
as determined by experimental testing. ................................................................................. 68

16
Figure 38. Moment-rotational-displacement curves for the glulam Nuki joint with B w of 1.5”, as
determined by experimental testing. ...................................................................................... 69
Figure 39. Summary of rotational stiffnesses of all prototypes of the glulam Nuki joint with B w of 1.5”,
as determined by experimental testing. ................................................................................. 70
Figure 40. Moment-rotational-displacement curves for the glulam Nuki joint for all beam widths, as
determined by experimental testing, with outliers removed. .................................................. 71
Figure 41. Moment-rotational-displacement curve for the glulam Nuki joint determined by the FEA
model. .................................................................................................................................... 72
Figure 42. Comparison of moment-rotational-displacement curves using different methods for the
glulam Nuki joint with a B w of 1”. ........................................................................................... 73
Figure 43. Comparison of moment-rotational-displacement curves using different methods for the
glulam Nuki joint with a B w of 1.5”. ........................................................................................ 73
Figure 44. Comparison of rotational stiffnesses of the glulam Nuki joint as determined by different
methods, with (a) B w of 1” and (b) B w of 1.5”. ......................................................................... 74
Figure 45. Comparison of embedment in the glulam Nuki joint as determined by (a) experimental
testing, (b) analytical model, and (c) FEA model. Images by Daniel Landez, Jan Brütting, and
Aliz Fischer; figure assembled by Aliz Fischer. ........................................................................ 76
Figure 46. Final beam designs for varying spans using different connection types. ....................... 83
Figure 47. Graphic to assist with Nuki beam and joint design, charting the rotational stiffness of Nuki
joints of varying widths b and depths h in the range 1 < h/b < 4. These values are calculated
for a column depth (dimension C d, based on terminology of Section 3) of 20 inches. ............ 88
Figure 48. Embodied carbon comparison between gravity frame design methods. Embodied carbon
scales for steel sections are abbreviated in order to keep scales practical for other methods. For
beam hangers, only fasteners into the primary member are illustrated; fasteners into the
secondary member are still included in quantities calculations and are reflected in the displayed
embodied carbon quantities. .................................................................................................. 97
Figure 49. Embodied carbon against span for different design methods. The open dot acknowledges
that a beam hanger product from a different line, the MEGANT series with more aluminum than
the Ricon series, was used for the 20 ft span. ......................................................................... 98
Figure 50. Preliminary assessment on the use of Nuki joints with rotational restraint for lateral systems,
elevation view. 30 psf wind loads were applied as point loads assuming a 10 ft tributary width,
as well as unfactored gravity loads outlined in Section 5.2. Timber columns measuring 20” x 20”
were modeled. .......................................................................................................................100
Figure 51. Summary of findings comparing the embodied carbon of a beam supported by
conventional beam hangers to that of a beam supported by Nuki joints. In the 20 ft span
pictured, embodied carbon savings of 76% are observed. ....................................................104
Figure 52. Specimen 1 of the Douglas fir Nuki joint of B w = 1", scanned after rotational stiffness
testing. ..................................................................................................................................122
Figure 53. Specimen 2 of the Douglas fir Nuki joint of B w = 1", scanned after rotational stiffness
testing. ..................................................................................................................................123
Figure 54. Specimen 3 of the Douglas fir Nuki joint of B w = 1", scanned after rotational stiffness
testing. ..................................................................................................................................124

17
Figure 55. Specimen 1 of the Douglas fir Nuki joint of B w = 1.5", scanned after rotational stiffness
testing. ..................................................................................................................................125
Figure 56. Specimen 2 of the Douglas fir Nuki joint of B w = 1.5", scanned after rotational stiffness
testing. ..................................................................................................................................126
Figure 57. Specimen 3 of the Douglas fir Nuki joint of B w = 1.5", scanned after rotational stiffness
testing. ..................................................................................................................................127
Figure 58. Specimen 5 of the glulam Nuki joint of B w = 1", scanned after rotational stiffness testing.
..............................................................................................................................................129
Figure 59. Specimen 6 of the glulam Nuki joint of B w = 1", scanned after rotational stiffness testing.
..............................................................................................................................................130
Figure 60. Specimen 2 of the glulam Nuki joint of B w = 1.5", scanned after rotational stiffness testing.
..............................................................................................................................................131
Figure 61. Comparison of uniformly loaded beam with varying boundary conditions. ..................134
Figure 62. Free body diagram, shear V(x), moment M(x), and deflection w(x) diagrams for the beam
supported with partial rotational restraint. ............................................................................135

18
List of tables
Table 1. Chronological overview of methods used in literature characterizing analysis and behavior
of joinery connections under static load. Shaded cells indicate strength characterization;
unshaded cells indicate stiffness characterization. .................................................................. 34
Table 2. Summary of Nuki joint case studies conducted. ............................................................... 55
Table 3. Nuki joint dimensions used in the case study. .................................................................. 56
Table 4. Material properties used for Douglas fir prototypes......................................................... 57
Table 5. Material properties used for glue-laminated timber prototypes. ...................................... 58
Table 6. Summary of rotational stiffnesses for the Douglas fir Nuki joint as determined by the
analytical model. ..................................................................................................................... 59
Table 7. Summary of mean rotational stiffnesses for the Douglas fir Nuki joint as determined by
experimental testing, with outliers removed. .......................................................................... 63
Table 8. Comparison of rotational stiffnesses of the Douglas fir Nuki joint as determined by different
methods. ................................................................................................................................ 65
Table 9. Comparison of percentage increases in stiffness between B w of 1" and B w of 1.5" of the
Douglas fir Nuki joint as determined by different methods. ................................................... 66
Table 10. Comparison of yield moment between models of the Douglas fir Nuki joint. ................ 66
Table 11. Summary of rotational stiffnesses for the glulam Nuki joint as determined by the analytical
model. .................................................................................................................................... 67
Table 12. Summary of mean rotational stiffnesses for the glulam Nuki joint as determined by
experimental testing, with outliers removed. .......................................................................... 71
Table 13. Summary of rotational stiffnesses for the glulam Nuki joint as determined by the FEA
model. .................................................................................................................................... 72
Table 14. Comparison of rotational stiffnesses of the glulam Nuki joint as determined by different
methods. ................................................................................................................................ 75
Table 15. Comparison of percentage increases in stiffness between B w of 1" and B w of 1.5" of the
glulam Nuki joint as determined by different methods. .......................................................... 75
Table 16. Comparison of yield moment between models of the glulam Nuki joint. ....................... 75
Table 17. Assumptions for comparative study................................................................................ 80
Table 18. Steel profiles selected for each span. ............................................................................. 84
Table 19. Limit states for the design of a timber beam supported at both ends with a connection of
shear capacity Vallow. ................................................................................................................ 84
Table 20. Beam hangers selected for each span based on shear demand. Beam hanger product
design values provided in MTC Solutions catalog (MTC Solutions 2020), and S-P-F glulam was
assumed. ................................................................................................................................ 85
Table 21. Selected beam sections for each span based on beam hanger selection. ...................... 86
Table 22. Key values for beam design of varying spans supported by beam hangers. .................. 86
Table 23. Limit states for the design of a timber beam supported at both ends with a Nuki joint of
known rotational stiffness k. .................................................................................................... 87
19
Table 24. Beam cross sections satisfying all limit states when supported by Nuki joints. Cross sections
are selected by picking a width b and finding the minimum depth h that satisfies all limit states,
within the range 1 < h/b < 4. .................................................................................................. 90
Table 25. Selected beam sections for each span based on limit states for beams supported by Nuki
joints, and associated rotational stiffness in the elastic region. .............................................. 90
Table 26. Key values for beam design of varying spans supported by Nuki joints. ........................ 91
Table 27. Material assumptions for calculation of embodied carbon. Embodied carbon coefficients
were obtained from Jones and Hammond 2019. .................................................................... 92
Table 28. Quantities of different gravity frame designs for a 10 ft span......................................... 93
Table 29. Quantities of different gravity frame designs for a 15 ft span......................................... 94
Table 30. Quantities of different gravity frame designs for a 20 ft span......................................... 95
Table 31. Quantities of different gravity frame designs for a 25 ft span......................................... 96
Table 32. Ratios between embodied carbon values for different spans and methods. .................. 98
Table 33. Quantities calculation for a Ricon S VS 140x60 beam hanger........................................140
Table 34. Quantities calculation for a Ricon S VS 200x80 beam hanger........................................141
Table 35. Aluminum quantities calculation for a MEGANT 310x100 beam hanger. ......................142
Table 36. Steel wire rod quantities calculation for a MEGANT 310x100 beam hanger. ................142
Table 37. Quantities calculation for a Ricon XL 390x80 beam hanger. ..........................................143

20
1 Introduction
1.1 Motivation
In a rapidly urbanizing world where buildings contribute to 40% of global carbon emissions,
structural designers are becoming increasingly conscientious of their roles and responsibilities in
the profession (International Energy Agency (IEA) and Global Alliance for Buildings and
Construction (GlobalABC) 2018). These responsibilities are acknowledged by professional
organizations who have challenged designers to reduce and eventually eliminate embodied
carbon by 2050 (“Structural Engineers 2050 Challenge” n.d.) or even 2030 (“Architecture 2030”
n.d.).
One major strategy for reducing embodied carbon in buildings is to select structural materials
with lower embodied carbon. As a result, structural timber is witnessing a recent revival in
modern construction. “Mass timber” is the term used for engineered wood products, glued-up
pieces of small timber, that have recently enabled large construction with embodied carbon at
a fraction of that of a concrete or steel building of the same scale.
Because most mass timber elements, like cross-laminated timber (CLT) and glue-laminated
timber (glulam), are standardized linear or planar elements, one trend in the mass timber
 Demi Fang 1 Introduction
SMBT 2020

movement is the use of metallic fasteners to connect timber elements. These range from nails
and screws to connection systems with more components and bearing plates (Figure 1, Figure
2, Figure 3, Figure 4, Figure 5).

Figure 1. Steel connections between glulam elements at the Bullitt Center (Miller Hull Partnership;
Seattle, WA, USA; 2013). Left: during construction (Court 2012), photo by John Stamets; right: interior
after completion (Schuchart n.d.).

Figure 2. Interior of the Promega Feynman Center “The Crossroads” manufacturing facility (Uihlein-
Wilson Architects, Inc.; Madison, WI, USA; 2013). (WoodWorks n.d.)
22
 Demi Fang 1 Introduction
SMBT 2020

Figure 3. Structural connection detail at T3 (Michael Green Architecture and DLR Group; Minneapolis,
MN, USA; 2016). (Brownell 2016)

Figure 4. Structural detailing of the Brock Commons Tallwood House (Acton Ostry Architects; Vancouver,
BC, Canada; 2017) at University of British Columbia. (naturally:wood 2016)

23
 Demi Fang 1 Introduction
SMBT 2020

Figure 5. Connection details of zipper trusses at the John W. Olver Design Building (Leers Weinzapfel
Associates; Amherst, MA, USA; 2017) at University of Massachusetts Amherst. (Miller 2017)

Even visually, these metallic connection components stand out, seemingly in contrast with the
original motivations of using timber as a more sustainable structural material. Their modern-day
use may have evolved out of both practicality and evolution of the screws and nails used in light-
frame timber construction. However, even before the invention of screws and nails, historic
timber construction across the world made use of joinery, or geometrically interlocking timber
joints.
1.2 Timber joinery in history
Joinery has been known to be implemented in historic timber construction in East Asia (Figure 6
and Figure 7), North America (Figure 9 and Figure 10), present-day United Kingdom (Figure 8),
Europe, and Egypt (Foliente 2000; Zwerger 2015).

24
 Demi Fang 1 Introduction
SMBT 2020

Figure 6. Bracket set joinery at Nandaimon, or the Great South Gate (Nara, Japan; 12th century) of Todai-
ji. Photo by Deanna MacDonald.

Figure 7. Joinery types in Chinese traditional timber architecture. (Shiping 1991)

25
Demi Fang 1 Introduction
SMBT 2020

Figure 8. Cruck framing, an English carpentry technique. Leigh Court Barn is the largest cruck framed
structure in Britain (Worcester, England; 1325). Photo by Simon Webb.

Figure 9. A modern example of North American timber framing tradition. Photo by Vermont Timber
Works Inc.
26
 Demi Fang 1 Introduction
SMBT 2020

Figure 10. Example joinery from the North American timber framing tradition. (Benson 1997)

Japan’s tradition of timber construction is particularly known for combining timber joinery with a
“design for deconstruction” approach that enabled structures like the Ise Jingu Shrine (4 BC) to
be rebuilt every 20 years as part of Shinto belief in death and renewal (Henrichsen and Bauer
2004).
1.3 Timber joinery in modern structures
As much as joinery was prevalent in historic construction, it is relatively rare in modern timber
construction. With its absence in building code, it only appears in relatively boutique
architectural structures. Some of these deliberately showcase and celebrate the use of the joinery
(Figure 11, Figure 12, Figure 13), while others are deliberate revivals of the historic timber
framing tradition at the residential scale (Figure 14).

27
Demi Fang 1 Introduction
SMBT 2020

Figure 11. Joinery detail at the Yusuhara Wooden Bridge Museum (Kengo Kuma & Associates; Yusuhara,
Japan; 2011). Photo by Takumi Ota.

Figure 12. All-timber structural system of the Tamedia Office Building (Shigeru Ban Architects; Zurich,
Switzerland; 2013). Photo by Didier Boy de la Tour.

28
Demi Fang 1 Introduction
SMBT 2020

Figure 13. "Cat's paw" timber tension joinery detail at the Writers Theatre (Studio Gang; Glencoe, IL,
USA; 2016). Photo by Steve Hall © Hedrich Blessing.

29
 Demi Fang 1 Introduction
SMBT 2020

Figure 14. Timber construction companies like Bensonwood create modern homes, like this Vermont
lake house (Ludlow, VT, USA; 2008) in the traditional style of North American timber framing. Photo by
Al Karevy.

The historic transition from joinery to fasteners arose out of the efficiency afforded by fasteners
to join standardized timber elements, compared to the time and labor required to carve joinery
geometries out of each timber element. However, as the above projects demonstrate, the
advancing role of digital fabrication in modern construction may tip the scale.
1.4 Potential benefits and opportunities
Introducing joinery into modern timber construction offers several benefits and opportunities. A
few are listed below, although not all are explored in this thesis.
1.4.1 Non-destructive assembly and Design for Deconstruction
The Ise Jingu Shrine establishes a historic precedent for using joinery for a Design for
Deconstruction (DfD) approach, a sustainable approach that extends the longevity of structures.
Though not explicitly addressed in this thesis, there is potential for the modern-day use of joinery
to reinvigorate aspects of DfD.
Admittedly, metallic fasteners that are conventionally used with timber elements in construction
systems of all scales may already be deconstruction-friendly, so the relative ease for DfD of
joinery connections should be evaluated.
The potential for joinery to be applied to DfD is related to its characteristic of non-destructive
assembly, which is not commonly seen in the construction industry. The logic of non-destructive
assembly is more frequently celebrated in products like furniture (Figure 15). On its product page,
“The Bed” from Thuma lists several marketable advantages of their product: that the Japanese-

30
 Demi Fang 1 Introduction
SMBT 2020

joinery-inspired design is “timeless, naturally noise-minimizing and easy to put together, while
eliminating the need for excess metal hardware.” Its speed of assembly is also highlighted.

Figure 15. "The Bed" from Thuma is marketed for its ease of assembly marked by an absence of metal
hardware. (“The Bed” n.d.)

1.4.2 Seismic systems

Many have observed the resilience of historic East Asian timber structures to major seismic events.
The structures’ resilience can be broadly attributed to a combination of factors: 1) the
redundancy offered by the structural configuration of stacking of multiple timber elements, 2)
the energy dissipation enabled by the plastic crushing of wood fibers that gives “softness” to
the global structure, and 3) the ability to replace damaged elements. While the potential for
joinery in modern-day seismic systems is not explicitly addressed in this thesis, a literature review
of modern analyses is provided in Section 2.2.
1.4.3 Reduced embodied carbon
By now, the sustainability benefits of timber as a structural material over conventional materials
like steel and concrete are well established (Gustavsson, Pingoud, and Sathre 2006; Lenzen and
Treloar 2002; Sathre and O’Connor 2010). Comparative studies like Stern 2018 account for the
tradeoff between embodied carbon and material strength by selecting timber over steel in planar
trusses. However, those tradeoffs have been studied on the structural design and quantification
of primary elements (truss elements) rather than their connections. Extending this tradeoff
analysis into the design of connection components has the potential to make the sustainability
31
 Demi Fang 1 Introduction
SMBT 2020

evaluations of structural systems even more holistic. In this thesis, one such tradeoff analysis that
accounts for connection components is presented in Section 5.
1.4.4 Biophilia
Biophilic design recognizes the benefits of human connection with nature. The exposure of
natural materials, such as timber, in the interior of architectural spaces constitutes one attribute
of biophilic design called “indirect experience of nature” (Kellert and Calabrese 2015). These
visual connections can reduce stress, improve cognitive performance, and boost moods in
humans working or living within those spaces (Browning, Ryan, and Clancy 2014). Depending on
the structural system, the visual effect of metallic connection details can be prominent in timber
structural systems (Figure 1, Figure 2, Figure 3, Figure 4, Figure 5). The principles of biophilic
design suggest that single-material timber systems (such as those depicted in Figure 11, Figure
12, Figure 13, and Figure 14) may offer a physiological benefit to inhabitants.
1.4.5 Digital fabrication
The mass timber construction industry enables more pre-fabricated elements, which has also
provided more opportunities for digital fabrication than traditionally used in the construction
industry. Custom CNC millwork can be done at minimal additional cost, ranging from duct
openings in CLT panels to milling mortises for joinery components, as modern-day timber
framing companies like Bensonwood have done. These technologies can additionally be
combined with other techniques like curved lamination or selective layups, challenging the
conventional standardization of rectilinear elements and offering new ways to achieve material
and structural efficiency. The comparative study presented in Section 5 envisions the potential
of combining these digital fabrication techniques to enable low-carbon structural systems.
1.5 Thesis aims
This thesis aims to investigate the feasibility of re-introducing timber joinery into large-scale
modern timber construction. How can their performance be quantified as structural joints? Can
they provide adequate structural rigidity and softness by design at the appropriate scales in
structural systems? Are there significant benefits in embodied carbon by using joinery
connections instead of conventional metallic connections? After a review of relevant literature, a
set of original contributions are set forth in this thesis to begin answering these questions.

32
2 Literature review
In this section, an overview of existing literature on the analysis of joinery connections in structural
systems is presented.
2.1 Static analysis
Previous studies have sought to characterize the behavior of timber joinery connections under
static load. Most methods can be classified under analytical, experimental, or numerical methods.
Analytical methods use wood mechanics to develop models with parameters to calculate
rotational stiffness of joinery connections. Experimental methods involve the creation of
prototypes and recording their performance under applied static loads. Numerical methods
involve the development of finite element analyses (FEA) where material properties and
geometries are input to computationally simulate the behavior of the joinery connection. Most
existing studies report and compare findings from at least two methods on a single joint type. A
summary of the existing literature is provided in Table 1.
Demi Fang 2 Literature review
SMBT 2020

Table 1. Chronological overview of methods used in literature characterizing analysis and behavior of
joinery connections under static load. Shaded cells indicate strength characterization; unshaded cells
indicate stiffness characterization.

Experimenta
Joint type Analytical Numerical
l
Pegged mortise-and-tenon,
(Brungraber
and full-size knee-braced X X X
1985)
timber bent frames
(Inayama 1991) None (theory of
X X -
(in Japanese) embedment)

(Schmidt and
Pegged mortise-and-tenon X X -
Mackay 1997)
(Erdil, Kasal, and Mortise and tenon (furniture
X X -
Eckelman 2005) scale)
Pegged mortise-and-tenon
(Erikson 2003) - X X
in knee-braced frames
(Chang, Hsu,
and Komatsu Continuous Nuki joint X X -
2006)
(Chang and Hsu Continuous Nuki joint,
X X -
2007) butted Nuki joint
(Lang and Fodor
Notched beams - X X
2007)
(Guan, Kitamori,
Nuki: beam through column
and Komatsu - X X
with wedges
2008)
(Shanks and
Pegged mortise-and-tenon X X X
Walker 2009)
(Kitamori,
Kataoka, and None (theory of
Komatsu 2009) X X X
embedment)
(in Japanese)
Nuki: beam through
(Komatsu et al.
column, and mortise-and- X -* -*
2009)
tenon
(Sebera and
Dovetail - - X
Šimek 2010)

34
 Demi Fang 2 Literature review
SMBT 2020

(Tannert, Lam,
Rounded dovetail X X X
and Vallée 2010)
(Ogawa, Sasaki,
and Yamasaki Watari-ago: notched beams X X -
2015)
(Shope 2016) Double birdsmouth X X -
(Roche 2017) Timber plates X X X
(S.-Y. Yeo et al. Dieh-Dou: stacked bracket
X X -
2016) system

2.1.1 Analytical methods

[Link] Wood mechanics
Wood is not an isotropic material, so when defining its stiffness, the orientation to grain is
important. In this thesis, 𝐸𝐸0 represents the material’s Young’s modulus parallel to grain and 𝐸𝐸90
represents the material’s Young’s modulus perpendicular to grain. If a timber material has a
known 𝐸𝐸0 and 𝐸𝐸90 , the Young’s modulus at some angle 𝛽𝛽 to the grain can be defined by
Hankinson’s formula:

𝐸𝐸0 ∗ 𝐸𝐸90 (1)

𝐸𝐸(𝛽𝛽) = 𝑛𝑛
𝐸𝐸0 cos 𝛽𝛽 + 𝐸𝐸90 sin𝑛𝑛 𝛽𝛽
where 𝑛𝑛 = 2 (Hankinson 1921).
[Link] Strength-based analyses of joinery connections
A handful of the represented literature, shaded in Table 1, test for the strength rather than for
stiffness of joinery connections. Quantifying strength is most notably relevant for pegged tension
joinery. The analytical method provided for the pegged mortise-and-tenon in Schmidt and
Mackay (1997) uses the European Yield Model, which focuses on the failure modes of the joint
under tension load, all of which involve shearing of the peg. Studies like Shanks and Walker (2009)
further develop the analytical model for this pegged mortise-and-tenon tension joint by
considering elasto-plastic behavior and energy dissipation of the peg. Shope (2016) uses an
analytical method that is generally unique in the literature, applying the concept of stress fields
to estimating the strength of a double birdsmouth joint under compressive load. Hankinson’s
formula is used to represent the reduction of stiffness associated with wood grain direction.
[Link] Characterizing rotational stiffness of joinery connections using the embedment model
Primarily, though, the literature presented in Table 1 seek to characterize the semi-rigid behavior
of joinery connections, i.e. quantify the stiffness provided by these joint connections. The
analytical method presented by Brungraber (1985) represents the joinery connection as a semi-
rigid joint with three springs (one axial, one shear, and one rotational). Stiffness matrices are
developed to represent the behavior of the joints under static load.
Since then, most literature rely on the theoretical framework of embedment established by
Inayama (1991). While the original work is presented in Japanese, the theoretical framework can

35
 Demi Fang 2 Literature review
SMBT 2020

be extrapolated from various English-language publications that apply the methods, such as
Komatsu et al. (2009), Ogawa, Sasaki, and Yamasaki (2015), Yeo et al. (2016), and Tanahashi and
Suzuki (2020). The theory and its methods are implemented in Fundamental Theory of Timber
Engineering by the Architectural Institute of Japan (Architectural Institute of Japan (AIJ) 2010).
The theory has been successfully implemented not only on traditional linear wood elements but
also on joinery connections between timber plates: Roche (2017) implements the embedment
theory in chapters 5-7 in characterizing the semi-rigidity of joinery connecting thin structural
wood panels.
In considering the behavior of most joinery connections, the “triangular embedment with
exponential-shaped additional length” model established by Inayama has been shown, and is
often used, to capture the rotational stiffness provided by rotational partial compression of one
wood element against another. For more documentation of other moment-resisting embedment
models, the reader is referred to section 2.2 of Roche (2017), in addition to Tanahashi and Suzuki
(2020).
The methodology from Inayama’s embedment theory applied in this work is presented in Section
3.
2.1.2 Experimental methods
Nearly all of the literature summarized in Table 1 validate the presented models against
experimental testing. The experimental testing usually involves applying some increasing load
from one timber component on another and recording the corresponding displacement. Scopes
in the literature range from material level embedment behavior to global timber frame behavior,
and prototypes were usually built to the corresponding scale. For studies that examine a specific
joinery type, usually prototypes of the specific joinery geometry are fabricated and tested.
2.1.3 Numerical methods
Most numerical simulations of joinery behavior included in Table 1 utilize a type of finite element
analysis (FEA) called contact FEA, which accounts for the stress at contact interfaces. These
interactions are important to model since they capture the local material behavior at interfaces
and pose a greater technical challenge than the scope of FEA typically used in structural
engineering modeling.
2.1.4 Static behavior of mortise-and-tenon-type joinery
One of the most fundamental joinery geometries is the mortise-and-tenon joint, also called the
Nuki joint in Japanese (Figure 16).
Several pieces in literature examine the Nuki joint in different variations; North American
variations typically include pegs for shear resistance, while Japanese variations sometimes
include wedges to close any differential gaps between mortise and tenon.

36
 Demi Fang 2 Literature review
SMBT 2020

Figure 16. The traditional Japanese Nuki geometry as a mortise-and-tenon-type joinery connection for
beams and columns.

For those studies that characterized the stiffness of the Nuki joint, regardless of the method used
to characterize the stiffness behavior, most demonstrated or utilized a bilinear model (Komatsu
et al. 2009; Guan, Kitamori, and Komatsu 2008; Chang, Hsu, and Komatsu 2006; Chang and Hsu
2007). The joint would first exhibit some initial stiffness relating load with displacement, or
moment with angular displacement. After some threshold (predicted by an analytical model or
numerical simulation, or observed in experimental testing), the stiffness would soften to a second
plastic stiffness which occurs when the crushing of wood perpendicular to the grain enters the
plastic region.
2.2 Dynamic analysis
A number of works have studied the seismic behavior joinery connections, particularly of East
Asian joinery connections. These studies are dominated by quasi-static tests, both experimental
and numerical (D’Ayala and Tsai 2008; Li et al. 2015; Chen Zhiyong et al. 2016; L. Chen et al.
2017; J. Chen et al. 2018; Y. Wu, Song, and Li 2018) .
Experimental dynamic testing are also common, sometimes with global scale models and shake
tables (J. Chen et al. 2018; Y. J. Wu, Song, and Luo 2017; Xie et al. 2018; Suzuki and Maeno
2006; Y. Wu et al. 2018). (In one outstanding case, microtremor and free vibration tests were
done on a newly built traditional pagoda in full scale (Hanazato et al. 2004).) Most works
examined the dynamic properties of joints in partial context or in isolation (Li et al. 2015; L. Chen
et al. 2017; Xue, Xu, and Xia 2018; Fujita et al. 2000; S. Y. Yeo et al. 2016). Most of these studies
report on the hysteresis behavior, stiffness degradation, and energy dissipation of the joints. For
global models, modal analyses were sometimes included (Hanazato et al. 2004; Y. Wu, Song,
and Ventura 2019).
For frames with mortise-and-tenon joints and even dovetail variants, pull-out is the most common
form of failure (Li et al. 2015; L. Chen et al. 2017). This failure was also observed directly in the
field, in a post-disaster investigation following the Tangshan earthquake of 1976. The same study
observed that in general, the timber structures in the affected area sustained much less damage
and collapse than did masonry structures (Huixian et al. 2002).
With shake table tests, cyclic loading is sometimes at a slow enough scale that it is still a quasi-
static test. Dynamic tests include microtremors/white noise (high frequency, small amplitude),
free vibration, sine wave, and earthquake history inputs.

37
 Demi Fang 2 Literature review
SMBT 2020

Aside from mortise-and-tenon and dovetail joints, the seismic behavior of stacked bracket sets
on tops of columns are also of interest. These joinery connections are sometimes reported to
demonstrate a higher damping than would be seen in the connections of reinforced concrete or
steel structures (Xie et al. 2018; Fujita et al. 2000; S. Y. Yeo et al. 2016).
In most studies, the effect of vertical loading was examined. Depending on the joint, vertical
loading could sometimes improve the joint stiffness, but increased vertical loading also meant
greater mass for greater inertial effects under earthquake loading (Chang et al. 2012; Chen
Zhiyong et al. 2016; S. Y. Yeo et al. 2016; J. Chen et al. 2018).
A few studies looked at the effect of tightness of the joinery connection on behavior. Generally,
the looser the joint, the lower the rotational stiffness (Xue, Xu, and Xia 2018). In works like
(Hanazato et al. 2004), the tightness in the joinery connections were cited as the likely reason
behind a much higher observed damping than predicted. Other works like (S. Y. Yeo et al. 2016)
identified friction force between contact surfaces as critical for the frame’s overall structural
integrity. The aforementioned common failure mode of tenon pull-out also indicates that friction
and tightness in joinery connections play an important role on failure.
A few authors have recognized the importance of the semi-rigidity of joints on P-delta effects in
the global structure and specifically addressed those effects (Li et al. 2015; King, Yen, and Yen
1996).
Generally, the good seismic behavior of structures with joinery connections are attributed to the
plastic deformation – and resulting energy dissipation and damping – that occur at the joinery
connections under seismic events (Chen Zhiyong et al. 2016; Tanabashi 1960; Que et al. 2017).
2.3 Digital fabrication and joinery
Recent developments have demonstrated the potential of robotic fabrication methods to
reproduce joinery connections, ranging from specific joints like the dovetail and finger joints
(Page 2017) to historic timber frame component replacement in Chile (Böhme, Zapata, and
Ansaldo 2017), to a complete and faithful reproduction of a traditional Japanese pagoda
(Takabayashi, Kado, and Hirasawa 2019). In practice, modern fabricators of timber frame
components are already using state-of-the-art CNC machines to produce their parts.
2.4 Evaluating sustainability of early-stage structural design exploration using
embodied carbon
The mitigation of carbon in the built environment is becoming more and more acknowledged as
a priority, particularly the mitigation of embodied carbon. Pomponi and Moncaster 2016 have
determined that pluralistic, multidisciplinary approaches are essential for conducting effective
life cycle analyses (LCAs) for improved methods of designing low-carbon structures, while De
Wolf 2017 identifies the two main pathways as (1) reducing structural material quantities and (2)
designing with materials with lower embodied carbon. Databases such as the Database of
Embodied Quantity Outputs (DEQO) represent the development of transparent, open-source,
and reliable resources to assist the design of low-carbon structures (C. De Wolf et al. 2020). Other
tools have arisen in industry that reflect a widespread desire to facilitate low-carbon structural
design (KT Innovations, thinkstep, and Autodesk 2019; Building Transparency 2019; CORE
Studio 2020).

38
 Demi Fang 2 Literature review
SMBT 2020

In parallel to these efforts to collect data is the effort to develop improved methodologies for
low-carbon structural design. One early study implementing LCA in early-stage structural design
demonstrated a 65.1% reduction in embodied energy and 67.2% reduction in carbon dioxide
emissions by selecting a cable long-span structural system over a truss system (Trussoni et al.
2015). In another study, this methodology was generalized for spanning trusses of varying
geometry and material combinations, examining and optimizing the tradeoff of timber truss
elements being lower in embodied carbon but also lower in strength (Stern 2018). In both of
these studies, primary structural elements were considered and the material contribution of
connection details were not included.
Another study assesses the feasibility of timber for structural systems of tall buildings, comparing
the embodied carbon of a tall timber tower to that of a benchmark concrete building. A hybrid
system is proposed where concrete joints provide rotational restraint to reduce timber floor plate
material. The all-timber scheme considered in Appendix C of this report is assumed to have an
all-pinned scheme and is found to be less materially efficient than the timber-concrete composite
system for a tall building (Skidmore, Owings & Merrill, LLP 2013).
2.5 Problem statement
Both the literature on the dynamic performance of joinery and the digital fabrication capabilities
available today clarify the potential for joinery connections to re-exist in modern timber
construction. However, the sustainability benefits of re-introducing joinery connections remain
unclear.
The literature sets out established methods for evaluating the stiffness of joinery connections.
The Nuki joint, or mortise-and-tenon beam-through-column joint, is identified as a fundamental
joinery geometry that has been evaluated and used for different structural systems.
However, the behavior of the Nuki joint has yet to be characterized in a way that is compatible
with and applicable to structural design at the scale of modern mass timber projects. At this scale,
it is also unclear how the Nuki joint compares to modern conventional timber connections. Both
of these gaps may explain why joinery connections are largely absent from modern-day mass
timber structures.
2.6 Thesis format
This thesis seeks to consolidate existing literature on characterizing the Nuki joint to improve the
accessibility of joinery analysis for structural designers. A comparative study of structural design
with joinery and conventional timber connections is presented to highlight the benefits of
applying joinery in modern timber construction. Both approaches aim to reduce the barrier of
entry of joinery to modern structural design.
First, models that have been developed in literature are reconstructed on a Nuki joint. The
analytical, experimental, and numerical methodologies described in Section 3 are reproducible
on other joinery geometries.
The results of applying these three analysis methods on the Nuki joint of varying beam widths is
presented in Section 4. The bilinear behavior described in literature is observed, and the models
are calibrated based on specific material properties of the experimental prototypes. Particular
attention is given to the elastic rotational stiffness and the threshold moment before the joint
exhibits plastic behavior, as those two parameters are to be applied in structural design.

39
 Demi Fang 2 Literature review
SMBT 2020

Finally, the analytical model developed in Section 4 is applied to a comparative study between
a beam supported by Nuki joints with a more conventional timber connection design (Section 5).
In particular, the embodied carbon quantities between the systems are compared.

40
3 Methodology for characterizing the static
mechanical behavior of the Nuki joint
In this section, the methodology for characterizing the static mechanical behavior of the Nuki
joint is described. These methods are based on literature and are used to determine the results
in Section 4.
3.1 Analytical method: general approach
As discussed in Section [Link], the “triangular embedment with exponential-shaped additional
length” model established by Inayama (1991) will primarily be used here to capture the rotational
stiffness provided by rotational partial compression of one wood element against another. The
triangular embedment refers to the compression of wood fibers perpendicular to the grain due
to the contact rotation of another element. The additional indirect embedment stress is
experienced by the exposed portion just outside the direct contact region. If the embedment is
acting in the z-direction, this additional indirect stress can occur in both x- and y-directions of
 Demi Fang 3 Methodology for characterizing the static
SMBT 2020 mechanical behavior of the Nuki joint

the compressed element (Figure 17). As the name of the method suggests, the indirect
embedment shapes can be approximated as an exponential function.

Figure 17. These illustrations show the surface deformation of the indirectly loaded area just outside of
the area under direct contact with embedment. (c) through (e) illustrate the case of rotational partial
compression from different angles. Figure by (Roche 2017), synthesizing images from (Inayama 1991;
1993; Architectural Institute of Japan (AIJ) 2010).

The indirect embedment shape takes the form of

𝑓𝑓(𝑥𝑥) = ∆𝑒𝑒 −𝛼𝛼𝛼𝛼 (2)

where ∆ represents the linear compression depth at the end of direct triangular embedment and
𝛼𝛼 represents the decay coefficient, a value that has been tested extensively in the literature but
is taken by the AIJ to be 1.5/𝑍𝑍0 , where 𝑍𝑍0 is the initial depth of the compressed element (Figure
17). Further documentation on literature investigating decay coefficients is provided in Section
2.2 of Roche 2017 (Roche 2017) under the heading “Complementary studies on the decay
coefficient ae”.
Integrating direct embedment ∆(𝑥𝑥) and indirect embedment 𝑓𝑓(𝑥𝑥) along x- and y-axes, the
values of compressed volumes 𝑉𝑉 can be obtained. Using Hooke’s law, compressed volumes can
then be used to calculate resultant contact force 𝑁𝑁:

42
 Demi Fang 3 Methodology for characterizing the static
SMBT 2020 mechanical behavior of the Nuki joint

𝑉𝑉 (3)
N=σA=εEA=(εA)E= 𝐸𝐸
𝑍𝑍0

Rewriting the equation as a function of rotation angle 𝜃𝜃:

𝑉𝑉 (𝜃𝜃 ) (4)
N(𝜃𝜃)= 𝐸𝐸(𝜃𝜃)
𝑍𝑍0 (𝜃𝜃 )

A yield depth ∆𝑦𝑦 = 𝜀𝜀𝑦𝑦 𝑍𝑍0 indicates that the crushing of wood has entered the plastic region, where
𝜀𝜀𝑦𝑦 represents the yield strain material property across the grain. For embedment depths beyond
∆𝑦𝑦 , a reduced material Young’s modulus 𝐸𝐸 must be used.
Once contact forces 𝑁𝑁 and the associated lever arm lengths 𝑎𝑎 are determined, the relationship
between moment 𝑀𝑀 and rotation angle 𝜃𝜃 can be determined. Summarized for elastic and plastic
regions, the relationship can be stated as:

M(θ) = � 𝑁𝑁𝑖𝑖 (𝜃𝜃 ) 𝑎𝑎𝑖𝑖 (𝜃𝜃 ) (5)

3.2 Analytical method: Nuki joint

This section details the analytical model specifically developed for the Nuki joint. The model
presented in this section is based on, and updated from, work presented in Fang, Mueller, et al.
2019.
When the beam is loaded, the Nuki joint experiences rotational displacement. This causes
rotational partial compression at two locations in the beam.
If the beam sufficiently pierced through the column, the behavior at both embedment locations
might be equivalent due to symmetry. However, due to limitations in the experimental testing
phase, this case study was asymmetric: on one end, the beam element was only inserted 1 inch
past the edge of the column element. As a result, the additional embedment length on that end
of the beam is limited by the amount of beam length remaining. The embedment at the two
locations are thus not equivalent in this particular case study and need to be calculated
separately. The subscript e is used to denote the ended beam end, while the subscript c is used
to denote the continuous beam end. The following discussion of parameters apply to each
embedment location.

43
 Demi Fang 3 Methodology for characterizing the static
SMBT 2020 mechanical behavior of the Nuki joint

(a) (b) (c)

θ beam element θy θy
θ
Bd
column element

Le
grain

θ θy
M
Cd
θy θ
L Cd/2

Bd/2 Lc
Δ
L Δ Ve,pl Vd,pl
yield limit
L = Cd / (2 cos θ) Δy
Ve,el Vd,el θ Ve,el Vd,el θ
Δ = L ∙ sin θ
Le ai L p,e Lp

Figure 18. Analytical model (as a cut section down the middle of the prototype width): (a) joint geometry
dimensions, (b) elastically compressed volumes, and (c) compressed volumes for an elastoplastic material
behavior. Elastically compressed volumes are shown in blue tones; plastic volumes are given in magenta
tones. The boxed diagrams in (b) and (c) represent the contact at the ended beam end and thus the
subscripts e are used; the diagrams are analogous for the continuous beam end, where subscripts e
would be replaced with subscripts c. Note that the formula for L has been updated since the original
published study. Adapted from original illustration by collaborator Jan Brütting.
𝜋𝜋
The following relationships apply for rotational displacement 𝜃𝜃 between 0 and radians,
2
measured relative to the center of rotation ⊗, which is here assumed to be fixed throughout.
Final results are presented only for some range of radians that approximately matches the range
observed in experimental testing for this case study.
𝑍𝑍 is used to denote the depth of material experiencing embedment; in this prototype, 𝑍𝑍 is
expressed as a function of 𝜃𝜃:

Z(𝜃𝜃) = 𝐵𝐵𝑑𝑑 cos 𝜃𝜃 (6)

∆ represents the embedment depth, the maximum embedment which occurs at the edge of
direct contact and additional embedment. It is a function of 𝜃𝜃:

∆(𝜃𝜃) = 𝐿𝐿(𝜃𝜃) sin 𝜃𝜃 (7)

The length 𝐿𝐿 represents the beam length experiencing embedment. Initially, this value is half the
𝐶𝐶
column depth, or 𝑑𝑑. As rotational displacement 𝜃𝜃 increases, so does 𝐿𝐿:
2

𝐶𝐶𝑑𝑑 /2 (8)
L(𝜃𝜃) =
cos 𝜃𝜃

44
 Demi Fang 3 Methodology for characterizing the static
SMBT 2020 mechanical behavior of the Nuki joint

∆𝑦𝑦 represents the embedment depth at the point of yielding, and 𝜃𝜃𝑦𝑦 represents the rotational
displacement at the point of yielding. Given the material yield strain 𝜀𝜀𝑦𝑦 across the grain, ∆𝑦𝑦 can
be calculated as

∆𝑦𝑦 = 𝜀𝜀𝑦𝑦 𝐵𝐵𝑑𝑑 (9)

Evaluating ∆𝑦𝑦 = ∆(𝜃𝜃𝑦𝑦 ), it follows from Equations (7) and (8) that

𝜀𝜀𝑦𝑦 𝐵𝐵𝑑𝑑 (10)

𝜃𝜃𝑦𝑦 = tan−1 � �
𝐶𝐶𝑑𝑑 /2

Once the yield limits ∆𝑦𝑦 and 𝜃𝜃𝑦𝑦 are determined, it is possible to assign the respective elastic and
plastic Young’s moduli to the respective compressed volumes. The compressed volumes can in
this case be calculated by calculating the area in section and multiplying by the uniform depth
𝐵𝐵𝑤𝑤 .
3.2.1 Elastic embedment
Figure 18(b) illustrates the elastic case where ∆< ∆𝑦𝑦 ; both compressed volumes are in the elastic
region, hence the blue tones. The lighter blue volume is under direct contact and is thus called
𝑉𝑉𝑑𝑑,𝑒𝑒𝑒𝑒 . The darker blue volume is the additional exponential-shaped length; the end that is limited
by the ended beam end is called 𝑉𝑉𝑒𝑒,𝑒𝑒𝑒𝑒 and the analogous compressed volume along the
continuous beam end is called 𝑉𝑉𝑐𝑐,𝑒𝑒𝑒𝑒 .
The exponential-shaped length takes the form of

𝑓𝑓(𝜃𝜃, 𝑥𝑥) = ∆(𝜃𝜃) ∗ 𝑒𝑒 −𝛼𝛼𝛼𝛼 (11)

where the 𝑥𝑥-axis lies along the beam edge, starting from the edge of embedment and positive
direction pointing away from direct contact. The decay coefficient 𝛼𝛼 is taken to be 6.5/𝐵𝐵𝑑𝑑 , a
value recommended in most literature. The literature also reports recommended embedment
lengths across which to integrate 𝑓𝑓(𝑥𝑥) in order to calculate the compressed volumes for
additional embedment 𝑉𝑉𝑒𝑒,𝑒𝑒𝑒𝑒 and 𝑉𝑉𝑐𝑐,𝑒𝑒𝑒𝑒 . Embedment length for the ended beam end, 𝐿𝐿𝑒𝑒 , is taken
to be 1 inch, whereas embedment length for the continuous beam end, 𝐿𝐿𝑒𝑒 , is taken from
literature to be 1.5𝐵𝐵𝑑𝑑 .
All compressed volumes for the elastic case (∆< ∆𝑦𝑦 ) can now be calculated:

1 (12)
𝑉𝑉𝑑𝑑,𝑒𝑒𝑒𝑒 (𝜃𝜃) = 𝐵𝐵𝑤𝑤 ∗ ∆(𝜃𝜃) ∗ 𝐿𝐿 cos 𝜃𝜃
2
𝐿𝐿𝑒𝑒
𝑉𝑉𝑒𝑒,𝑒𝑒𝑒𝑒 (𝜃𝜃) = 𝐵𝐵𝑤𝑤 ∗ � 𝑓𝑓(𝜃𝜃, 𝑥𝑥) 𝑑𝑑𝑑𝑑 , 𝐿𝐿𝑒𝑒 = 1 in. (13)
𝑥𝑥=0

𝐿𝐿𝑐𝑐
𝑉𝑉𝑐𝑐,𝑒𝑒𝑒𝑒 (𝜃𝜃) = 𝐵𝐵𝑤𝑤 ∗ � 𝑓𝑓(𝜃𝜃, 𝑥𝑥) 𝑑𝑑𝑑𝑑 , 𝐿𝐿𝑐𝑐 = 1.5𝐵𝐵𝑑𝑑 (14)
𝑥𝑥=0

45
 Demi Fang 3 Methodology for characterizing the static
SMBT 2020 mechanical behavior of the Nuki joint

The compressed volumes can then be converted into resultant contact forces using the
𝑉𝑉(𝜃𝜃)
relationship N(𝜃𝜃)= ( ) 𝐸𝐸(𝜃𝜃) . While the compressed volume is initially being compressed
𝑍𝑍 𝜃𝜃
precisely perpendicular to the grain, as 𝜃𝜃 increases, the wood grain becomes more and more
aligned with the compression direction. 𝐸𝐸 is thus taken as a function of 𝜃𝜃 using the following
expression of Hankinson’s formula:

𝐸𝐸0 ∗ 𝐸𝐸90
𝐸𝐸(𝜃𝜃) = 𝜋𝜋 𝜋𝜋 , 𝑛𝑛 = 2 (15)
𝐸𝐸0 cos𝑛𝑛 � − 𝜃𝜃� + 𝐸𝐸90 sin𝑛𝑛 � − 𝜃𝜃�
2 2

The resultant contact forces are:

𝑉𝑉𝑑𝑑,𝑒𝑒𝑒𝑒 (𝜃𝜃) (16)

𝑁𝑁𝑑𝑑,𝑒𝑒𝑒𝑒 (𝜃𝜃) = 𝐸𝐸(𝜃𝜃)
𝑍𝑍(𝜃𝜃)
𝑉𝑉𝑒𝑒,𝑒𝑒𝑒𝑒 (𝜃𝜃) (17)
𝑁𝑁𝑒𝑒,𝑒𝑒𝑒𝑒 (𝜃𝜃) = 𝐸𝐸(𝜃𝜃)
𝑍𝑍(𝜃𝜃)
𝑉𝑉𝑐𝑐,𝑒𝑒𝑒𝑒 (𝜃𝜃) (18)
𝑁𝑁𝑐𝑐,𝑒𝑒𝑒𝑒 (𝜃𝜃) = 𝐸𝐸(𝜃𝜃)
𝑍𝑍(𝜃𝜃)

Finally, the moment arm for each resultant force needs to be calculated to determine the
rotational stiffness. The moment arm is the distance (in the direction perpendicular to the
associated resultant force) between the center of rotation ⊗ and the centroid of each
compressed volume.

2
𝑎𝑎𝑑𝑑,𝑒𝑒𝑒𝑒 (𝜃𝜃) = 𝐿𝐿 (𝜃𝜃) ∗ cos 𝜃𝜃
3 (19)
𝐶𝐶𝑑𝑑
Note that 𝐿𝐿(𝜃𝜃) ∗ cos 𝜃𝜃 = 2
, so 𝑎𝑎𝑑𝑑,𝑒𝑒𝑒𝑒 is actually a constant and not a function of 𝜃𝜃.

𝐿𝐿𝑒𝑒
𝐶𝐶𝑑𝑑 1 (20)
𝑎𝑎𝑒𝑒,𝑒𝑒𝑒𝑒 (𝜃𝜃) = + � 𝑥𝑥 ∗ 𝑓𝑓(𝜃𝜃, 𝑥𝑥)𝑑𝑑𝑑𝑑, 𝐿𝐿𝑒𝑒 = 1 in.
2 𝑉𝑉𝑒𝑒,𝑒𝑒𝑒𝑒 𝑥𝑥=0
𝐿𝐿𝑐𝑐
𝐶𝐶𝑑𝑑 1 (21)
𝑎𝑎𝑐𝑐,𝑒𝑒𝑒𝑒 (𝜃𝜃) = + � 𝑥𝑥 ∗ 𝑓𝑓(𝜃𝜃, 𝑥𝑥)𝑑𝑑𝑑𝑑, 𝐿𝐿𝑐𝑐 = 1.5𝐵𝐵𝑑𝑑
2 𝑉𝑉𝑐𝑐,𝑒𝑒𝑒𝑒 𝑥𝑥=0

3.2.2 Plastic embedment

Figure 18(c) illustrates the case where ∆> ∆𝑦𝑦 ; magenta tones denote the compressed volumes
that are operating in the plastic region. These volumes must be calculated separately from the
elastic ones.

46
 Demi Fang 3 Methodology for characterizing the static
SMBT 2020 mechanical behavior of the Nuki joint

While calculating compressed volumes, it is helpful to define a few dimensions. First, 𝐿𝐿𝑝𝑝 is
defined as the length along the axis of the beam, between elastic and plastic compressed
volumes (Figure 18):

𝐶𝐶𝑑𝑑 ∆𝑦𝑦 (22)

𝐿𝐿𝑝𝑝 (𝜃𝜃) = ∗ �1 − �
2 ∆(𝜃𝜃)

Additionally, 𝐿𝐿𝑝𝑝,𝑒𝑒 and 𝐿𝐿𝑝𝑝,𝑐𝑐 are defined as the lengths along the x-axis of additional embedment
length (for their respective beam ends) at which the additional embedment shape matches yield
embedment depth ∆𝑦𝑦 :

𝐿𝐿𝑝𝑝,𝑒𝑒 = 𝐿𝐿𝑝𝑝,𝑐𝑐 = 𝑓𝑓 −1 (∆𝑦𝑦 ) (23)

The compressed volumes can then be defined as the following, with some volumes split into two
parts for ease of centroid calculation:

1 (24)
𝑉𝑉𝑑𝑑,𝑒𝑒𝑒𝑒,1 (𝜃𝜃) = 𝐵𝐵𝑤𝑤 ∗ ∆𝑦𝑦 ∗ �𝐿𝐿(𝜃𝜃) ∗ cos 𝜃𝜃 − 𝐿𝐿𝑝𝑝 (𝜃𝜃)�
2
𝑉𝑉𝑑𝑑,𝑒𝑒𝑒𝑒,2 (𝜃𝜃) = 𝐵𝐵𝑤𝑤 ∗ ∆𝑦𝑦 𝐿𝐿𝑝𝑝 (𝜃𝜃) (25)

1 (26)
𝑉𝑉𝑑𝑑,𝑝𝑝𝑝𝑝 (𝜃𝜃) = 𝐵𝐵𝑤𝑤 ∗ �∆(𝜃𝜃) − ∆𝑦𝑦 �𝐿𝐿𝑝𝑝 (𝜃𝜃)
2

The compressed volumes at the ended beam end are:

𝑉𝑉𝑒𝑒,𝑒𝑒𝑒𝑒,1 = 𝐵𝐵𝑤𝑤 ∗ 𝐿𝐿𝑝𝑝,𝑒𝑒 ∆𝑦𝑦 (27)

𝐿𝐿𝑒𝑒
𝑉𝑉𝑒𝑒,𝑒𝑒𝑒𝑒,2 (𝜃𝜃) = 𝐵𝐵𝑤𝑤 ∗ � 𝑓𝑓(𝜃𝜃, 𝑥𝑥) 𝑑𝑑𝑑𝑑 (28)
𝐿𝐿𝑝𝑝,𝑒𝑒

𝐿𝐿𝑝𝑝,𝑒𝑒
𝑉𝑉𝑒𝑒,𝑝𝑝𝑝𝑝 (𝜃𝜃) = 𝐵𝐵𝑤𝑤 ∗ � 𝑓𝑓(𝜃𝜃, 𝑥𝑥) − ∆𝑦𝑦 𝑑𝑑𝑑𝑑 (29)
0

Analogously, for the continuous beam end:

𝑉𝑉𝑐𝑐,𝑒𝑒𝑒𝑒,1 = 𝐵𝐵𝑤𝑤 ∗ 𝐿𝐿𝑝𝑝,𝑐𝑐 ∆𝑦𝑦 (30)

𝐿𝐿𝑐𝑐
𝑉𝑉𝑐𝑐,𝑒𝑒𝑒𝑒,2 (𝜃𝜃) = 𝐵𝐵𝑤𝑤 ∗ � 𝑓𝑓(𝜃𝜃, 𝑥𝑥) 𝑑𝑑𝑑𝑑 (31)
𝐿𝐿𝑝𝑝,𝑐𝑐

𝐿𝐿𝑝𝑝,𝑐𝑐
𝑉𝑉𝑐𝑐,𝑝𝑝𝑝𝑝 (𝜃𝜃) = 𝐵𝐵𝑤𝑤 ∗ � 𝑓𝑓(𝜃𝜃, 𝑥𝑥) − ∆𝑦𝑦 𝑑𝑑𝑑𝑑 (32)
0

47
 Demi Fang 3 Methodology for characterizing the static
SMBT 2020 mechanical behavior of the Nuki joint

In the next step of converting the compressed volumes into resultant contact forces, again the
𝑉𝑉(𝜃𝜃)
relationship N(𝜃𝜃)= 𝐸𝐸(𝜃𝜃) is applied. However, the Young’s moduli for plastically compressed
𝑍𝑍(𝜃𝜃)
volumes must be adjusted to reflect the plastic stiffness. A bilinear material model is assumed
where 𝑃𝑃𝑅𝑅90 represents the factor by which the Young’s modulus in the elastic region is reduced
to obtain the Young’s modulus in the plastic region.

𝑉𝑉𝑑𝑑,𝑒𝑒𝑒𝑒,𝑖𝑖 (𝜃𝜃) (33)

𝑁𝑁𝑑𝑑,𝑒𝑒𝑒𝑒,𝑖𝑖 (𝜃𝜃) = 𝐸𝐸(𝜃𝜃), for all i
𝑍𝑍(𝜃𝜃)
𝑉𝑉𝑑𝑑,𝑝𝑝𝑝𝑝 (𝜃𝜃) (34)
𝑁𝑁𝑑𝑑,𝑝𝑝𝑝𝑝 (𝜃𝜃) = �𝑃𝑃𝑅𝑅90 ∗ 𝐸𝐸(𝜃𝜃)�
𝑍𝑍(𝜃𝜃)
𝑉𝑉𝑒𝑒,𝑒𝑒𝑒𝑒,𝑖𝑖 (𝜃𝜃) (35)
𝑁𝑁𝑒𝑒,𝑒𝑒𝑒𝑒,𝑖𝑖 (𝜃𝜃) = 𝐸𝐸(𝜃𝜃), for all i
𝑍𝑍(𝜃𝜃)
𝑉𝑉𝑒𝑒,𝑝𝑝𝑝𝑝 (𝜃𝜃) (36)
𝑁𝑁𝑒𝑒,𝑝𝑝𝑝𝑝 (𝜃𝜃) = �𝑃𝑃𝑅𝑅90 ∗ 𝐸𝐸(𝜃𝜃)�
𝑍𝑍(𝜃𝜃)
𝑉𝑉𝑐𝑐,𝑒𝑒𝑒𝑒,𝑖𝑖 (𝜃𝜃) (37)
𝑁𝑁𝑐𝑐,𝑒𝑒𝑒𝑒,𝑖𝑖 (𝜃𝜃) = 𝐸𝐸(𝜃𝜃), for all i
𝑍𝑍(𝜃𝜃)
𝑉𝑉𝑐𝑐,𝑝𝑝𝑝𝑝 (𝜃𝜃) (38)
𝑁𝑁𝑐𝑐,𝑝𝑝𝑝𝑝 (𝜃𝜃) = �𝑃𝑃𝑅𝑅90 ∗ 𝐸𝐸(𝜃𝜃)�
𝑍𝑍(𝜃𝜃)

Finally, moment arms are determined. For direct contact:

2 𝐶𝐶𝑑𝑑 (39)
𝑎𝑎𝑑𝑑,𝑒𝑒𝑒𝑒,1 (𝜃𝜃) = � − 𝐿𝐿𝑝𝑝 (𝜃𝜃)�
3 2
𝐶𝐶𝑑𝑑 𝐿𝐿𝑝𝑝 (𝜃𝜃) (40)
𝑎𝑎𝑑𝑑,𝑒𝑒𝑒𝑒,2 (𝜃𝜃) = −
2 2
𝐶𝐶𝑑𝑑 𝐿𝐿𝑝𝑝 (𝜃𝜃) (41)
𝑎𝑎𝑑𝑑,𝑝𝑝𝑝𝑝 (𝜃𝜃) = −
2 3

Moment arms at ended beam end:

𝐶𝐶𝑑𝑑 𝐿𝐿𝑝𝑝,𝑒𝑒 (42)

𝑎𝑎𝑒𝑒,𝑒𝑒𝑒𝑒,1 = +
2 2
𝐿𝐿𝑒𝑒
𝐶𝐶𝑑𝑑 1
𝑎𝑎𝑒𝑒,𝑒𝑒𝑒𝑒,2 (𝜃𝜃) = + � 𝑥𝑥 ∗ 𝑓𝑓(𝜃𝜃, 𝑥𝑥) 𝑑𝑑𝑑𝑑 (43)
2 𝑉𝑉𝑒𝑒.𝑒𝑒𝑒𝑒,2 𝐿𝐿𝑝𝑝,𝑒𝑒

𝐿𝐿𝑝𝑝,𝑒𝑒
𝐶𝐶𝑑𝑑 1 (44)
𝑎𝑎𝑒𝑒,𝑝𝑝𝑝𝑝 (𝜃𝜃) = + � 𝑥𝑥 ∗ (𝑓𝑓(𝜃𝜃, 𝑥𝑥) − ∆𝑦𝑦 ) 𝑑𝑑𝑑𝑑
2 𝑉𝑉𝑒𝑒,𝑝𝑝𝑝𝑝 0

48
 Demi Fang 3 Methodology for characterizing the static
SMBT 2020 mechanical behavior of the Nuki joint

Analogously, moment arms at continuous beam end:

𝐶𝐶𝑑𝑑 𝐿𝐿𝑐𝑐,𝑒𝑒 (45)

𝑎𝑎𝑐𝑐,𝑒𝑒𝑒𝑒,1 = +
2 2
𝐿𝐿𝑐𝑐
𝐶𝐶𝑑𝑑 1
𝑎𝑎𝑐𝑐,𝑒𝑒𝑒𝑒,2 (𝜃𝜃) = + � 𝑥𝑥 ∗ 𝑓𝑓(𝜃𝜃, 𝑥𝑥) 𝑑𝑑𝑑𝑑 (46)
2 𝑉𝑉𝑐𝑐.𝑒𝑒𝑒𝑒,2 𝐿𝐿𝑝𝑝,𝑐𝑐

𝐿𝐿𝑝𝑝,𝑐𝑐
𝐶𝐶𝑑𝑑 1 (47)
𝑎𝑎𝑐𝑐,𝑝𝑝𝑝𝑝 (𝜃𝜃) = + � 𝑥𝑥 ∗ (𝑓𝑓(𝜃𝜃, 𝑥𝑥) − ∆𝑦𝑦 ) 𝑑𝑑𝑑𝑑
2 𝑉𝑉𝑐𝑐,𝑝𝑝𝑝𝑝 0

3.2.3 Contribution of friction

Friction from wood-wood contact – that is, direct contact only – contributes to the moment
experienced at the joint. This moment contribution is calculated using the material’s static
coefficient of friction 𝜇𝜇. The force of friction 𝐹𝐹𝑓𝑓 acts along the length of the beam element and
arises from every normal force acting at both contact points:

𝐹𝐹𝑓𝑓 (𝜃𝜃) = 𝜇𝜇 ∗ � Ni (θ) (48)

i∈d

Note that the additional indirect embedment forces do not contribute to friction.
The associated moment arm for 𝐹𝐹𝑓𝑓 is notated as 𝑎𝑎𝑓𝑓 and represents the perpendicular distance
between the friction forces. In this case, it is equivalent to the beam depth:

𝑎𝑎𝑓𝑓 = 𝐵𝐵𝑑𝑑 (49)

3.2.4 Determining rotational stiffness

Generally, the resultant forces 𝑁𝑁 and associated moment arms 𝑎𝑎 can be synthesized as follows
to determine the relationship between moment and angular displacement:

𝑀𝑀(𝜃𝜃) = � 𝑁𝑁𝑖𝑖 (𝜃𝜃)𝑎𝑎𝑖𝑖 (𝜃𝜃) (50)

𝑖𝑖∈(𝑑𝑑∪𝑒𝑒∪𝑐𝑐∪𝑓𝑓)

where 𝑑𝑑 represents direct contact, 𝑒𝑒 represents additional embedment along the ended end, 𝑐𝑐
represents additional embedment along the continuous end, and 𝑓𝑓 represents friction.
Because the formulas for some normal forces are dependent on whether the plastic region is
reached or not, details of the summations for each region are detailed below.
If the rotations are in the elastic region (𝜃𝜃 < 𝜃𝜃𝑦𝑦 ), the relationship can be expressed as follows:

𝑀𝑀𝑒𝑒𝑒𝑒 (𝜃𝜃) = 2[𝑁𝑁𝑑𝑑,𝑒𝑒𝑒𝑒 (𝜃𝜃)𝑎𝑎𝑑𝑑,𝑒𝑒𝑒𝑒 (𝜃𝜃)]+𝑁𝑁𝑒𝑒,𝑒𝑒𝑒𝑒 (𝜃𝜃)𝑎𝑎𝑒𝑒,𝑒𝑒𝑒𝑒 (𝜃𝜃) + 𝑁𝑁𝑐𝑐,𝑒𝑒𝑒𝑒 (𝜃𝜃)𝑎𝑎𝑐𝑐,𝑒𝑒𝑒𝑒 (𝜃𝜃) (51)
+ 𝜇𝜇𝜇𝜇𝑑𝑑,𝑒𝑒𝑒𝑒 (𝜃𝜃)𝐵𝐵𝑑𝑑
49
 Demi Fang 3 Methodology for characterizing the static
SMBT 2020 mechanical behavior of the Nuki joint

where formulas for each are detailed in Section 3.2.1.

If in the plastic region (𝜃𝜃 ≥ 𝜃𝜃𝑦𝑦 ), the relationship in this case is:

𝑀𝑀𝑝𝑝𝑝𝑝 (𝜃𝜃) = 2[𝑁𝑁𝑑𝑑,𝑒𝑒𝑒𝑒 (𝜃𝜃)𝑎𝑎𝑑𝑑,𝑒𝑒𝑒𝑒 (𝜃𝜃) + 𝑁𝑁𝑑𝑑,𝑝𝑝𝑝𝑝 (𝜃𝜃)𝑎𝑎𝑑𝑑,𝑝𝑝𝑝𝑝 (𝜃𝜃)]+𝑁𝑁𝑒𝑒,𝑒𝑒𝑒𝑒,1 (𝜃𝜃)𝑎𝑎𝑒𝑒,𝑒𝑒𝑒𝑒,1 (𝜃𝜃)
+ 𝑁𝑁𝑒𝑒,𝑒𝑒𝑒𝑒,2 (𝜃𝜃)𝑎𝑎𝑒𝑒,𝑒𝑒𝑒𝑒,2 (𝜃𝜃) + 𝑁𝑁𝑐𝑐,𝑒𝑒𝑒𝑒,1 (𝜃𝜃)𝑎𝑎𝑐𝑐,𝑒𝑒𝑒𝑒,1 (𝜃𝜃) (52)
+ 𝑁𝑁𝑐𝑐,𝑒𝑒𝑒𝑒,2 (𝜃𝜃)𝑎𝑎𝑐𝑐,𝑒𝑒𝑒𝑒,2 (𝜃𝜃)+ 𝑁𝑁𝑒𝑒,𝑝𝑝𝑝𝑝 (𝜃𝜃)𝑎𝑎𝑒𝑒,𝑝𝑝𝑝𝑝 (𝜃𝜃) + 𝑁𝑁𝑐𝑐,𝑝𝑝𝑝𝑝 (𝜃𝜃)𝑎𝑎𝑐𝑐,𝑝𝑝𝑝𝑝 (𝜃𝜃)
+ 𝜇𝜇[𝑁𝑁𝑑𝑑,𝑒𝑒𝑒𝑒 (𝜃𝜃)𝐵𝐵𝑑𝑑 + 𝑁𝑁𝑑𝑑,𝑝𝑝𝑝𝑝 (𝜃𝜃)𝐵𝐵𝑑𝑑 ]

where formulas for each are detailed in Section 3.2.2.

The general formula can be written as a piecewise function of the functions in each region:

𝑀𝑀𝑒𝑒𝑒𝑒 (𝜃𝜃) when 𝜃𝜃 < 𝜃𝜃𝑦𝑦 (53)

𝑀𝑀(𝜃𝜃) = �
𝑀𝑀𝑝𝑝𝑝𝑝 (𝜃𝜃) when 𝜃𝜃 ≥ 𝜃𝜃𝑦𝑦

The resulting function can be plotted and compared to results obtained using experimental or
numerical methods.
A bilinear behavior is usually observed. The two slopes representing the two rotational stiffnesses
𝜕𝜕𝜕𝜕
associated with each plot can be determined by calculating 𝜕𝜕𝜕𝜕
at a point in each region. The
𝜃𝜃𝑦𝑦
point in the elastic region was selected to be at 𝜃𝜃 = and in the plastic region 𝜃𝜃 = 3𝜃𝜃𝑦𝑦 .
2

The code used to run the analytical models are provided in Appendix 1.
3.3 Experimental methods
Prototypes of each joint are fabricated and load-tested. Two fabrication methods are used in this
work; the first is hand carpentry, and the second is digital fabrication using 3-axis CNC machines.
Load-testing of the rotational stiffness of the joint is done by applying a force on the beam
element while anchoring the column element, as in Figure 19. The rotation at the joint can be
measured using string potentiometers.

50
 Demi Fang 3 Methodology for characterizing the static
SMBT 2020 mechanical behavior of the Nuki joint

Figure 19. Setup for experimental load-testing of Nuki joint prototypes. Graphic made with collaborator
Jan Brütting.

For the Douglas fir prototypes, 3 specimens for each beam width were fabricated using hand
carpentry techniques. For the glulam prototypes, 6 specimens for each beam width were digitally
fabricated with a robotic CNC router. The opening in the columns has rounded corners equal to
the size of router ball end, i.e. 3/8 inches (Figure 20).

Figure 20. Mortise with rounded corners in the digitally fabricated prototypes.

In the test setup, which is shown in Figure 19, the column element of the joint was fixed to the
table, while the beam element was attached to a hand-screw-actuated load cell of 1000 lb
capacity. The beam-column connection was not press-fit, and varied in tightness. The total gap
between the beam and column ranged from 0 to about 0.08 inches. A linear displacement was
applied at the beam end using the load cell, and the resulting rotation at the joint was measured
by two string potentiometers. Figure 21 shows example progressive frames from one prototype.

Figure 21. Progressive frames from testing rotational stiffness of glulam specimens.

3.4 Numerical methods

Any finite element analysis (FEA) program that has the capability of modeling the behavior of
contact interfaces could be used to model the behavior of joinery connections. In this study, LS-
DYNA was used where numerical methods were used. Equipped with material data provided by
material testing at MIT, collaborators at Arup San Francisco (Aliz Fischer, Benshun Shao, Julieta

51
 Demi Fang 3 Methodology for characterizing the static
SMBT 2020 mechanical behavior of the Nuki joint

Moradei) led the effort in creating material cards and numerical simulations of the joint under
the same loading conditions as in the experimental testing. A side-by-side comparison of
experimental and numerical testing setups is shown in Figure 22.

Figure 22. (a) Experimental setup for Nuki joint testing, and (b) boundary and loading conditions for
numerical simulation. Figure by collaborators Aliz Fischer and Jan Brütting.

Using material testing to determine material parameters of the glulam (Table 5), a numerical
model was built for the glulam prototypes by collaborators at Arup San Francisco. Other values,
such as Poisson ratio 𝜐𝜐 = 0.39, were estimated from literature (Forest Products Laboratory 2010).
Contact between beam and column is defined by numerical parameters representing the contact
stiffness, the friction between surfaces, and the initial gap size. In the model, the gap size
between the beam and the column was set up based on observation of the fabrication, with 0.5
mm gap each side of the beam. At the top and bottom faces the gap was smaller, 0.025 mm, to
ensure numerical stability. The analysis assumed no initial stress from joining the two pieces.

Figure 23. Stress in the beam perpendicular to grain (psi) throughout the numerical simulation for the Bw
= 1 in. specimen. Analysis performed by Aliz Fischer, Benshun Shao, and Nick Sherrow-Groves of Arup
San Francisco.

52
 Demi Fang 3 Methodology for characterizing the static
SMBT 2020 mechanical behavior of the Nuki joint

Selected images of the progression of the numerical simulation are shown in Figure 23. The
contour plot shows the stresses perpendicular to the grain of the beam. The simulation was
stopped when the analysis showed numerical instabilities as the damage at the contact became
excessive.
For more details on the FEA methods, the reader is referred to Shao et al. 2019 and Fang,
Moradei, et al. 2019.
3.5 Statement of contributions
While final findings and interpretations were synthesized by the author, parts of the
methodologies and results presented in Section 4 were made possible by a number of
collaborators.
Collaborators from the Arup San Francisco team (Aliz Fischer, Benshun Shao, Julieta Moradei,
Nick Sherrow-Groves) developed and ran numerical simulations (Section 4.3.3).
Collaborators from EPFL (Jan Brütting, Corentin Fivet) helped develop the theory (Section 3.2)
and code (Appendix 1) for the analytical model.
MIT undergraduate researcher Danny Landez carried out digital fabrication and experimental
testing of the glulam prototypes (Section 4.3.2) under conditions as close as possible to those
carried out by the author for the Douglas fir prototypes.
3.6 Summary
In this chapter, the analytical, experimental, and numerical approaches for characterizing the
rotational stiffness of the Nuki joint are synthesized and presented based on existing literature.
These methods are applied in Section 4 to characterize the rotational stiffness of a Nuki joint
across varying beam widths. The methodology outlined may be adapted to other joinery
geometries. A catalog detailing the rotational stiffnesses of different joinery geometries
validated by these methods may be instrumental in promoting the use of joinery in modern
structural design.

53
Demi Fang 3 Methodology for characterizing the static
SMBT 2020 mechanical behavior of the Nuki joint

54
4 Static mechanical behavior of the Nuki joint
In this chapter, three methods for determining the mechanical behavior of the Nuki joint is
described, and their results compared. The mechanical behavior studied is the rotational stiffness
of the joint, a property that could determine how important a role this joint might play as a
boundary condition in structural design.
Two sets of studies were done for the Nuki, differing in materials used and method of fabricating
prototypes. These studies and their respective analysis methods are summarized in Table 2.
Table 2. Summary of Nuki joint case studies conducted.

Nuki joint
Analytical Experimental Numerical
case studies
X
Douglas fir X -
(hand carpentry)
X
Glulam X X
(CNC)
 Demi Fang 4 Static mechanical behavior of the Nuki joint
SMBT 2020

The geometry of the Nuki joint is illustrated in Figure 24. The joint consists of a beam element
acting as a tenon piercing through a fully mortised column element. The beam element will be
loaded on one side to test the rotational stiffness of the joint. This loading condition will create
double-contact rotational partial compression.

Figure 24. Nuki joint and associated dimensions examined in the case study.

The specific dimensions selected for the case study are summarized in Table 3. Two beam width
values were investigated.
Table 3. Nuki joint dimensions used in the case study.

1
Column width 𝐶𝐶𝑤𝑤 = 3 in.
4
1
Column depth 𝐶𝐶𝑑𝑑 = 3 in.
4
1
Beam depth 𝐵𝐵𝑑𝑑 = 3 in.
4
𝐵𝐵𝑤𝑤 = 1 in.
Beam width
and 𝐵𝐵𝑤𝑤 = 1.5 in.

4.1 Material properties

Where possible, standards defined by ASTM D143-14 (ASTM International 2014) were followed
as closely as possible to test for material properties of the timber.
For Douglas fir, 𝐸𝐸0 and 𝐸𝐸90 moduli were determined by taking the slopes of 5 specimens for each
modulus, prepared according to ASTM D143-14 (ASTM International 2014). Because the test
described for compression perpendicular to grain did not represent a pure 𝐸𝐸90 value, the true
𝐸𝐸90 value was estimated to be 90% the tested average value. Yield strain 𝜀𝜀𝑦𝑦 was determined by
examining the testing data for compression perpendicular to grain and taking an average (Figure
25). The value for the coefficient of friction 𝜇𝜇 was selected according to typical values found in
literature. A summary of the material properties used for analysis of the Douglas fir prototypes is
provided in Table 4.

56
 Demi Fang 4 Static mechanical behavior of the Nuki joint
SMBT 2020

Figure 25. Generalized stress-strain behavior of timber under compression perpendicular to grain, and
associated material values.

Table 4. Material properties used for Douglas fir prototypes.

Property Value Source

Average of stress-strain
Stiffness (Young’s modulus)
𝐸𝐸0 = 1.57 × 106 lb/in.2 slopes recorded from 5
parallel to grain
specimens
90% of average of stress-
Stiffness (Young’s modulus)
𝐸𝐸90 = 36,500 lb/in.2 strain slopes recorded
perpendicular to grain
from 5 specimens
Yield strain Average value observed
compression perpendicular to 𝜀𝜀𝑦𝑦 = 0.018 from 5 specimens
grain
Average value observed
Stiffness plastic reduction factor 𝑃𝑃𝑅𝑅90 = 0.07
from 5 specimens
Typical value found in
Static coefficient of friction 𝜇𝜇 = 0.2
literature

For the glulam, 𝐸𝐸0 and 𝐸𝐸90 moduli were also determined by taking the slopes of 8 specimens for
each modulus, prepared according to ASTM D143-14 (ASTM International 2014). To determine
the pure 𝐸𝐸90 value, the compression-perpendicular-to-grain test was simulated in an FEA model
and the 𝐸𝐸90 value tuned to match the experimental testing results (in this case, a 70% reduction
of the tested value). Yield strain 𝜀𝜀𝑦𝑦 was determined by examining the testing data for
compression perpendicular to grain and taking an average. The value for the coefficient of
friction 𝜇𝜇 was selected according to typical values found in literature. A summary of the material
properties used for analysis of the glulam prototypes is provided in Table 5.

57
 Demi Fang 4 Static mechanical behavior of the Nuki joint
SMBT 2020

Table 5. Material properties used for glue-laminated timber prototypes.

Property Value Source

Average of stress-strain
Stiffness (Young’s modulus)
𝐸𝐸0 = 1.96 × 106 lb/in.2 slopes recorded from 5
parallel to grain
specimens
70% of average of stress-
Stiffness (Young’s modulus)
𝐸𝐸90 = 55,800 lb/in.2 strain slopes recorded
perpendicular to grain
from 5 specimens
Average value observed
Yield strain 𝜀𝜀𝑦𝑦 = 0.017
from 5 specimens
Average value observed
Stiffness plastic reduction factor 𝑃𝑃𝑅𝑅90 = 0.118
from 5 specimens
Typical value found in
Static coefficient of friction 𝜇𝜇 = 0.2
literature

4.2 Results: Douglas fir prototypes

The results for the prototypes made with Douglas fir are summarized below.
4.2.1 Analytical results
The 𝑀𝑀(𝜃𝜃) plots as determined by the analytical model of the Douglas fir Nuki joint are charted
in Figure 26.

Figure 26. Moment-rotational-displacement curve for the Douglas fir Nuki joint determined by the
analytical model.

58
 Demi Fang 4 Static mechanical behavior of the Nuki joint
SMBT 2020

The slopes in each linear region of elastic and plastic behavior are extracted and summarized in
Table 6.
Table 6. Summary of rotational stiffnesses for the Douglas fir Nuki joint as determined by the analytical
model.

Elastic stiffness Plastic stiffness

𝐵𝐵𝑤𝑤 = 1 in. 76,800 lb-in./rad 29,500 lb-in./rad
𝐵𝐵𝑤𝑤 = 1.5 in. 115,200 lb-in./rad 44,300 lb-in./rad
% increase in stiffness from
50% 50%
𝐵𝐵𝑤𝑤 = 1 in. to 𝐵𝐵𝑤𝑤 = 1.5 in.

𝜀𝜀𝑦𝑦 𝐵𝐵𝑑𝑑
Yield angle can be computed as 𝜃𝜃𝑦𝑦 = tan−1 � � according to the model established in Section
𝐶𝐶𝑑𝑑 /2
3.2, and yield moment 𝑀𝑀𝑦𝑦 is the corresponding moment at 𝜃𝜃 = 𝜃𝜃𝑦𝑦 . Yield moment is 2765 lb-in.
and 4147 lb-in. for 𝐵𝐵𝑤𝑤 = 1 in. and 𝐵𝐵𝑤𝑤 = 1.5 in., respectively.
4.2.2 Experimental results
Scans of the tested beam elements showing embedment are provided in Appendix 2.
The plots of moment against rotational displacement as extracted from experimental testing of
all 3 prototypes of the Douglas fir Nuki joint, 𝐵𝐵𝑤𝑤 = 1 in., are plotted in Figure 27.

Figure 27. Moment-rotational-displacement curves for the Douglas fir Nuki joint with Bw of 1”, as
determined by experimental testing.

59
 Demi Fang 4 Static mechanical behavior of the Nuki joint
SMBT 2020

The slopes in each linear region of elastic and plastic behavior for all prototypes with 𝐵𝐵𝑤𝑤 = 1 in.
are summarized in Figure 28.

Rotational stiffnesses
Bw = 1 in.
Rotational stiffness (lb-in./rad) 180000
160000
140000
120000
100000
80000
60000
40000
20000
0
1 2 3
Specimen Number

Elastic stiffness Plastic stiffness

Figure 28. Summary of rotational stiffnesses of all prototypes of the Douglas fir Nuki joint with Bw of 1”,
as determined by experimental testing.

The elastic stiffness of Specimen 1 is considered an outlier and is not included in future
calculations and averages.
Yield moment 𝑀𝑀𝑦𝑦 is approximated for each specimen by observation as the moment at which
the elastic stiffness begins to soften, rounded to the nearest 10 lb-in. (Figure 27). Specimen 1 is
estimated to have a 𝑀𝑀𝑦𝑦 of 2240 lb-in., while Specimens 2 and 3 are estimated to have a 𝑀𝑀𝑦𝑦 of
2670 lb-in. The 𝑀𝑀𝑦𝑦 of Specimen 1 is discarded as an outlier.
Next, the plots of moment against rotational displacement as extracted from experimental
testing of all 3 prototypes of the Douglas fir Nuki joint, 𝐵𝐵𝑤𝑤 = 1.5 in., are plotted in Figure 28.

60
 Demi Fang 4 Static mechanical behavior of the Nuki joint
SMBT 2020

Figure 29. Moment-rotational-displacement curves for the Douglas fir Nuki joint with Bw of 1.5”, as
determined by experimental testing.

The slopes in each linear region of elastic and plastic behavior for all prototypes with 𝐵𝐵𝑤𝑤 = 1.5 in.
are summarized in Figure 30.

Rotational stiffnesses
Bw = 1.5 in.
180000
Rotational stiffness (lb-in./rad)

160000
140000
120000
100000
80000
60000
40000
20000
0
1 2 3
Specimen Number

Elastic stiffness Plastic stiffness

Figure 30. Summary of rotational stiffnesses of all prototypes of the Douglas fir Nuki joint with Bw of 1.5”,
as determined by experimental testing.

61
 Demi Fang 4 Static mechanical behavior of the Nuki joint
SMBT 2020

Yield moment 𝑀𝑀𝑦𝑦 is approximated for each specimen by observation as the moment at which
the elastic stiffness begins to soften, rounded to the nearest 10 lb-in. (Figure 28). Specimens 1
and 3 are estimated to have a 𝑀𝑀𝑦𝑦 of 3530 lb-in., while Specimen 2 is estimated to have a 𝑀𝑀𝑦𝑦 of
2140 lb-in. The 𝑀𝑀𝑦𝑦 of Specimen 2 is discarded as an outlier.
The plots for moment against rotational displacement for all prototypes are combined in Figure
31.

Figure 31. Moment-rotational-displacement curves for the Douglas fir Nuki joint for all beam widths, as
determined by experimental testing.

Finally, averages of rotational stiffness values for both beam width dimensions are summarized
in Table 7.

62
 Demi Fang 4 Static mechanical behavior of the Nuki joint
SMBT 2020

Table 7. Summary of mean rotational stiffnesses for the Douglas fir Nuki joint as determined by
experimental testing, with outliers removed.

Elastic stiffness Plastic stiffness

Mean = 88,100 lb-in./rad Mean = 31,300 lb-in./rad
(Std dev = 7200 lb-in./rad) (Std dev = 5600 lb-in./rad)
𝐵𝐵𝑤𝑤 = 1 in.
Specimens 2, 3 of Specimens 1, 2, 3 of
𝐵𝐵𝑤𝑤 = 1 in. 𝐵𝐵𝑤𝑤 = 1 in.
Mean = 110,300 lb-in./rad 35,400 lb-in./rad
(Std dev = 4200 lb-in./rad) (Std dev = 4700 lb-in./rad)
𝐵𝐵𝑤𝑤 = 1.5 in.
Specimens 2, 3, 6 of Specimens 2, 6 of
𝐵𝐵𝑤𝑤 = 1.5 in. 𝐵𝐵𝑤𝑤 = 1.5 in.
% increase in stiffness from
25% 13%
𝐵𝐵𝑤𝑤 = 1 in. to 𝐵𝐵𝑤𝑤 = 1.5 in.

4.2.3 Comparison of results

The plots of moment against rotational displacement as determined by both analytical and
experimental methods for the Douglas fir Nuki joint are plotted in Figure 32 for 𝐵𝐵𝑤𝑤 = 1 in. and
in Figure 33 for 𝐵𝐵𝑤𝑤 = 1.5 in.

Figure 32. Comparison of moment-rotational-displacement curves using different methods for the
Douglas fir Nuki joint with a Bw of 1”.

63
 Demi Fang 4 Static mechanical behavior of the Nuki joint
SMBT 2020

Figure 33. Comparison of moment-rotational-displacement curves using different methods for the
Douglas fir Nuki joint with a Bw of 1.5”.

A comparison of the slopes in each linear region of elastic and plastic behavior is plotted in
Figure 34.

64
 Demi Fang 4 Static mechanical behavior of the Nuki joint
SMBT 2020

(a) Rotational stiffnesses (b) Rotational stiffnesses

Bw = 1" Bw = 1.5"
140000 140000

120000 120000

100000 100000
Rotational stiffness (lb-in./rad)

Rotational stiffness (lb-in./rad)

80000 80000

60000 60000

40000 40000

20000 20000

0 0
Experimental Analytical FEA Experimental Analytical FEA
mean w/o mean w/o
outliers outliers

Elastic stiffness Plastic stiffness Elastic stiffness Plastic stiffness

Figure 34. Comparison of rotational stiffnesses of the Douglas fir Nuki joint as determined by different
methods, with (a) Bw of 1” and (b) Bw of 1.5”. FEA was not carried out for these Douglas fir prototypes.

A numerical tabulation of these stiffnesses is presented in Table 8.

Table 8. Comparison of rotational stiffnesses of the Douglas fir Nuki joint as determined by different
methods.

𝐵𝐵𝑤𝑤 = 1 in. 𝐵𝐵𝑤𝑤 = 1.5 in.

Elastic stiffness Plastic stiffness Elastic stiffness Plastic stiffness
[lb-in./rad] [lb-in./rad] [lb-in./rad] [lb-in./rad]
Experimental
mean, outliers 88,100 31,300 110,300 35,400
removed
76,800 29,500 115,200 44,300
Analytical model
[-13%] [-6%] [+4%] [+25%]
[%] = % difference from experimental mean

65
 Demi Fang 4 Static mechanical behavior of the Nuki joint
SMBT 2020

A comparison of the percentage increases in stiffness from increasing beam width is presented
in Table 9.
Table 9. Comparison of percentage increases in stiffness between Bw of 1" and Bw of 1.5" of the Douglas
fir Nuki joint as determined by different methods.

Elastic stiffness Plastic stiffness

Experimental model 25% 13%
Analytical model 50% 50%

Finally, a comparison of 𝑀𝑀𝑀𝑀 values is presented in Table 10.

Table 10. Comparison of yield moment between models of the Douglas fir Nuki joint.

𝐵𝐵𝑤𝑤 = 1 in. 𝐵𝐵𝑤𝑤 = 1.5 in.

Experimental mean, outliers
2670 lb-in. 3530 lb-in.
removed
2765 lb-in. 4147 lb-in.
Analytical model
[+4%] [+17%]
[%] = % difference from experimental mean

4.3 Results: glulam prototypes

The results for the prototypes made with glulam are summarized below.
4.3.1 Analytical results
The 𝑀𝑀(𝜃𝜃) plots as determined by the analytical model of the glulam Nuki joint are charted in
Figure 35.

66
 Demi Fang 4 Static mechanical behavior of the Nuki joint
SMBT 2020

Figure 35. Moment-rotational-displacement curve for the glulam Nuki joint determined by the analytical
model.

The slopes in each linear region of elastic and plastic behavior are extracted and summarized in
Table 11.
Table 11. Summary of rotational stiffnesses for the glulam Nuki joint as determined by the analytical
model.

Elastic stiffness Plastic stiffness

𝐵𝐵𝑤𝑤 = 1 in. 127,300 lb-in./rad 52,700 lb-in./rad
𝐵𝐵𝑤𝑤 = 1.5 in. 191,000 lb-in./rad 79,000 lb-in./rad
% increase in stiffness from
50% 50%
𝐵𝐵𝑤𝑤 = 1 in. to 𝐵𝐵𝑤𝑤 = 1.5 in.

𝜀𝜀𝑦𝑦 𝐵𝐵𝑑𝑑
Yield angle can be computed as 𝜃𝜃𝑦𝑦 = tan−1 � 𝐶𝐶 � according to the model established in Section
𝑑𝑑 /2
3.2, and yield moment 𝑀𝑀𝑦𝑦 is the corresponding moment at 𝜃𝜃 = 𝜃𝜃𝑦𝑦 . Yield moment is 4328 lb-in.
and 6494 lb-in. for 𝐵𝐵𝑤𝑤 = 1 in. and 𝐵𝐵𝑤𝑤 = 1.5 in., respectively.
4.3.2 Experimental results
Scans of the tested beam elements showing embedment are provided in Appendix 2.
The plots of moment against rotational displacement as extracted from experimental testing of
all 6 prototypes of the glulam Nuki joint, 𝐵𝐵𝑤𝑤 = 1 in., are plotted in Figure 36.

67
 Demi Fang 4 Static mechanical behavior of the Nuki joint
SMBT 2020

Figure 36. Moment-rotational-displacement curves for the glulam Nuki joint with Bw of 1”, as determined
by experimental testing.

The slopes in each linear region of elastic and plastic behavior for all prototypes with 𝐵𝐵𝑤𝑤 = 1 in.
are summarized in Figure 37.

Rotational stiffnesses
Bw = 1 in.
300000
Rotational stiffness (lb-in./rad)

250000

200000

150000

100000

50000

0
1 2 3 4 5 6
Specimen Number

Elastic stiffness Plastic stiffness

Figure 37. Summary of rotational stiffnesses of all prototypes of the glulam Nuki joint with Bw of 1”, as
determined by experimental testing.

Specimens 3 and 4 (dashed lines in Figure 36) exhibit abnormal behavior. They are considered
outliers and are not considered for the rest of the study.
68
 Demi Fang 4 Static mechanical behavior of the Nuki joint
SMBT 2020

Yield moment 𝑀𝑀𝑦𝑦 is approximated for each specimen by observation as the moment at which
the elastic stiffness begins to soften, rounded to the nearest 10 lb-in. (Figure 36). All specimens
remaining after removing outliers (Specimens 1, 2, 5, 6) are estimated to have a 𝑀𝑀𝑦𝑦 of 3400 lb-
in.
The plots of moment against rotational displacement as extracted from experimental testing of
all 6 prototypes of the glulam Nuki joint, 𝐵𝐵𝑤𝑤 = 1.5 in., are plotted in Figure 38.

Figure 38. Moment-rotational-displacement curves for the glulam Nuki joint with Bw of 1.5”, as
determined by experimental testing.

The slopes in each linear region of elastic and plastic behavior for all prototypes with 𝐵𝐵𝑤𝑤 = 1.5 in.
are summarized in Figure 39.

69
 Demi Fang 4 Static mechanical behavior of the Nuki joint
SMBT 2020

Rotational stiffnesses
Bw = 1.5 in.
400000

Rotational stiffness (lb-in./rad)

350000
300000
250000
200000
150000
100000
50000
0
1 2 3 4 5 6
Specimen Number

Elastic stiffness Plastic stiffness

Figure 39. Summary of rotational stiffnesses of all prototypes of the glulam Nuki joint with Bw of 1.5”, as
determined by experimental testing.

Specimens 1, 4, and 5 qualitatively exhibit abnormal behavior (dashed lines in Figure 38). They
are considered outliers and are not considered for the rest of the study. Furthermore, it can be
seen that the loading of Specimen 3 was stopped before reaching a secondary stiffness, so its
initially reported Stiffness 2 value is not used.
Yield moment 𝑀𝑀𝑦𝑦 is approximated for each specimen by observation as the moment at which
the elastic stiffness begins to soften, rounded to the nearest 10 lb-in. (Figure 38). Specimens 2
and 6 are estimated to have a 𝑀𝑀𝑦𝑦 of 4350 lb-in., and Specimen 3 is estimated to have a 𝑀𝑀𝑦𝑦 of
5240 lb-in. The average 𝑀𝑀𝑦𝑦 of these specimens is 4650 lb-in.
The plots for moment against rotational displacement for all beam widths are combined in Figure
40.

70
 Demi Fang 4 Static mechanical behavior of the Nuki joint
SMBT 2020

Figure 40. Moment-rotational-displacement curves for the glulam Nuki joint for all beam widths, as
determined by experimental testing, with outliers removed.

Finally, averages of rotational stiffness values for both beam width dimensions are summarized
in Table 12.
Table 12. Summary of mean rotational stiffnesses for the glulam Nuki joint as determined by
experimental testing, with outliers removed.

Elastic stiffness Plastic stiffness

Mean = 141,600 lb-in./rad 35,500 lb-in./rad
(Std dev = 7900 lb-in./rad) (Std dev = 2400 lb-in./rad)
𝐵𝐵𝑤𝑤 = 1 in.
Specimens 1, 2, 5, 6 of Specimens 1, 2, 5, 6 of
𝐵𝐵𝑤𝑤 = 1 in. 𝐵𝐵𝑤𝑤 = 1 in.
267,100 lb-in./rad 64,400 lb-in./rad
(Std dev = 21,900 lb-in./rad) (Std dev = 16,100 lb-in./rad)
𝐵𝐵𝑤𝑤 = 1.5 in.
Specimens 2, 3, 6 of Specimens 2, 6 of
𝐵𝐵𝑤𝑤 = 1.5 in. 𝐵𝐵𝑤𝑤 = 1.5 in.
% increase in stiffness from
89% 81%
𝐵𝐵𝑤𝑤 = 1 in. to 𝐵𝐵𝑤𝑤 = 1.5 in.

4.3.3 Numerical results

The 𝑀𝑀(𝜃𝜃) plots as determined by the FEA model of the glulam Nuki joint are charted in Figure
41.
71
 Demi Fang 4 Static mechanical behavior of the Nuki joint
SMBT 2020

Figure 41. Moment-rotational-displacement curve for the glulam Nuki joint determined by the FEA
model.

The slopes in each linear region of elastic and plastic behavior are extracted and summarized in
Table 13.
Table 13. Summary of rotational stiffnesses for the glulam Nuki joint as determined by the FEA model.

Elastic stiffness Plastic stiffness

𝐵𝐵𝑤𝑤 = 1 in. 166,800 lb-in./rad 45,200 lb-in./rad
𝐵𝐵𝑤𝑤 = 1.5 in. 248,900 lb-in./rad 59,700 lb-in./rad
% increase in stiffness from
49% 32%
𝐵𝐵𝑤𝑤 = 1 in. to 𝐵𝐵𝑤𝑤 = 1.5 in.

Yield moment 𝑀𝑀𝑦𝑦 is approximated by observation as the moment at which the elastic stiffness
begins to soften, rounded to the nearest 10 lb-in. (Figure 41). 𝑀𝑀𝑦𝑦 is estimated to be 2940 lb-in.
and 4400 lb-in. for 𝐵𝐵𝑤𝑤 = 1 in. and 𝐵𝐵𝑤𝑤 = 1.5 in., respectively.
4.3.4 Comparison of results
The plots of moment against rotational displacement as determined by both analytical and
experimental methods for the glulam Nuki joint are plotted in Figure 42 for 𝐵𝐵𝑤𝑤 = 1 in. and in
Figure 43 for 𝐵𝐵𝑤𝑤 = 1.5 in.

72
Demi Fang 4 Static mechanical behavior of the Nuki joint
SMBT 2020

Figure 42. Comparison of moment-rotational-displacement curves using different methods for the
glulam Nuki joint with a Bw of 1”.

Figure 43. Comparison of moment-rotational-displacement curves using different methods for the
glulam Nuki joint with a Bw of 1.5”.

73
 Demi Fang 4 Static mechanical behavior of the Nuki joint
SMBT 2020

A comparison of the slopes in each linear region of elastic and plastic behavior is plotted in
Figure 44.

(a) Rotational stiffnesses (b) Rotational stiffnesses

Bw = 1" Bw = 1.5"
300000 300000

250000 250000
Rotational stiffness (lb-in./rad)

Rotational stiffness (lb-in./rad)

200000 200000

150000 150000

100000 100000

50000 50000

0 0
Experimental Analytical FEA Experimental Analytical FEA
mean w/o mean w/o
outliers outliers

Elastic stiffness Plastic stiffness Elastic stiffness Plastic stiffness

Figure 44. Comparison of rotational stiffnesses of the glulam Nuki joint as determined by different
methods, with (a) Bw of 1” and (b) Bw of 1.5”.

A numerical tabulation of these stiffnesses is presented in Table 14.

74
 Demi Fang 4 Static mechanical behavior of the Nuki joint
SMBT 2020

Table 14. Comparison of rotational stiffnesses of the glulam Nuki joint as determined by different
methods.

𝐵𝐵𝑤𝑤 = 1 in. 𝐵𝐵𝑤𝑤 = 1.5 in.

Elastic stiffness Plastic stiffness Elastic stiffness Plastic stiffness
[lb-in./rad] [lb-in./rad] [lb-in./rad] [lb-in./rad]
Experimental
mean, outliers 141,600 35,500 267,100 64,400
removed
127,300 52,700 191,000 79,000
Analytical model
[-10%] [+48%] [-29%] [+23%]

Numerical 166,800 45,200 248,900 59,700

model [+18%] [+27%] [-7%] [-7%]
[%] = % difference from experimental mean

A comparison of the percentage increases in stiffness from increasing beam width is presented
in Table 15.
Table 15. Comparison of percentage increases in stiffness between Bw of 1" and Bw of 1.5" of the glulam
Nuki joint as determined by different methods.

Elastic stiffness Plastic stiffness

Experimental model 89% 81%
Analytical model 50% 50%
Numerical model 49% 32%

A comparison of 𝑀𝑀𝑦𝑦 values is presented in Table 16.

Table 16. Comparison of yield moment between models of the glulam Nuki joint.

𝐵𝐵𝑤𝑤 = 1 in. 𝐵𝐵𝑤𝑤 = 1.5 in.

Experimental mean, outliers
3400 lb-in. 4647 lb-in.
removed
4328 lb-in. 6494 lb-in.
Analytical model
[+27%] [+40%]
2940 lb-in. 4400 lb-in.
Numerical model
[-14%] [-5%]
[%] = % difference from experimental mean

75
 Demi Fang 4 Static mechanical behavior of the Nuki joint
SMBT 2020

A qualitative comparison of embedment among all three methods for the glulam Nuki joint is
shown in Figure 45.

Figure 45. Comparison of embedment in the glulam Nuki joint as determined by (a) experimental
testing, (b) analytical model, and (c) FEA model. Images by Daniel Landez, Jan Brütting, and Aliz Fischer;
figure assembled by Aliz Fischer.

4.4 Discussion
4.4.1 Comparison of overall moment-rotational-displacement behavior
With the exception of outlier experimental prototypes which were not considered, the results
from all models generally demonstrated the expected bilinear behavior: an initial elastic stiffness
followed by a softening to a secondary stiffness in their plots of moment against rotational
displacement (Figure 32, Figure 33, Figure 42, Figure 43).
In Figure 43, the numerical model shows an unexpected kink in the plastic region around a
rotational displacement of about 0.045 radians. Collaborator Benshun Shao reasons that this kink
occurs when the governing normal contact force shifts from the beam element against the inside
face of the mortise to the beam element against the outside face of the column; in other words,
it occurs when the beam element rotates enough to “turn the corner” of the mortise. One
possible explanation for why the transition is not smoother might be that the mesh size is not
fine enough.
The analytical model was deliberately selected to capture embedment with additional
exponential-shaped length as shown in Figure 45 (b). This behavior was evident in both
experimental testing and FEA testing (Figure 45 (a) and (c), respectively).
4.4.2 Comparison of stiffness values
Table 8 and Table 14 show the quantitative comparisons of stiffness across different methods. In
the Douglas fir models (Table 8), the analytical model usually predicted within 20% of
experimental values, except for plastic stiffness for 𝐵𝐵𝑤𝑤 = 1.5 in. which would not typically be used
in design. In the glulam models (Table 14), the analytical model tended to underestimate the
elastic stiffness demonstrated in the experimental models and overestimate the plastic stiffness
demonstrated in the experimental models. The numerical model’s stiffness values were within
30% of experimental stiffness values.

76
 Demi Fang 4 Static mechanical behavior of the Nuki joint
SMBT 2020

Possible explanations for the higher discrepancies due to limitations in experimental testing are
also discussed in Section 4.4.5.
4.4.3 Comparison of relative increases in rotational stiffness from increasing beam width
Table 9 and Table 15 show the percentage increases in stiffness demonstrated by increasing
beam width by 50% (from 1 in. to 1.5 in.). The expected value, 50%, is reflected in the analytical
model.
The experimental prototypes in Douglas fir do not show as much of a stiffness increase (Table
9), but this phenomenon is an acknowledged limitation of the experimental testing results and
is discussed more in Section 4.4.5.
On the other hand, in the glulam experimental prototypes (Table 15), the mean elastic stiffness
and mean plastic stiffness increase by 80-90%. This is an unexpectedly high increase and suggest
that experimental testing could be improved by implementing the suggestions discussed in
Section 4.4.5.
The FEA models for the glulam prototypes demonstrate an increase in stiffness of 49% for elastic
stiffness but only an increase of 32% for plastic stiffness due to an increase in beam width (Table
15). The kink mentioned in Section 4.4.1 may have affected the perceived measure of the plastic
stiffness; improving the model to smooth out the kink might produce a relationship between
plastic stiffness and beam width closer to expected.
4.4.4 Comparison of yield moment values
In the Douglas fir joints, the analytical model predicted a yield moment within 20% of the
experimental values (Table 10). In the glulam joints, the analytical model’s predictions were
farther off, over-predicting the yield moment by 30-40% of the experimental values (Table 16).
Because 𝑀𝑀𝑦𝑦 predictions from the analytical model would be used for design, it is important to
note this pattern of over-prediction to apply adequate safety factors.
4.4.5 Limitations in experimental testing
The most evident limiting factor was the quantity and quality of specimens examined for
experimental testing. For example, in the Douglas fir prototypes (for which 3 prototypes were
fabricated and tested for each beam width), it is difficult to identify the differences between
beam widths when all moment-rotational-displacement curves are plotted together as in Figure
31. After throwing out outliers, only two prototypes were considered for some stiffness values,
which may have contributed to a high standard deviation for elastic stiffness of 𝐵𝐵𝑤𝑤 = 1 in. (Table
7).
For the glulam studies, twice as many prototypes for each beam width were fabricated and tested
(6 each), but again unexpected testing behavior resulted in necessarily removing several outlier
specimens for each beam width. After removing outliers, the behavior and stiffnesses of the
remaining 𝐵𝐵𝑤𝑤 = 1 in. prototypes seemed to be fairly consistent (Table 12), but the behavior and
stiffnesses for the remaining 𝐵𝐵𝑤𝑤 = 1.5 in. prototypes were more sparse and varied (Table 12).
The quantity of specimens produced and tested were limited by time and resources.
Possible explanations for abnormal testing behavior include 1) variations in the material as timber
is a heterogeneous material, 2) changes in humidity and temperature between fabrication and
testing that may have changed the moisture content in the timber. Both possibilities can lead to

77
 Demi Fang 4 Static mechanical behavior of the Nuki joint
SMBT 2020

variations in tightness of fit between beam and column elements. Based on the presence of
outlier behavior in both the hand-crafted Douglas fir prototypes and CNC-ed glulam prototypes,
it is likely that these variations in fit may not necessarily be alleviated by digital fabrication. It is
worth noting that a variation of the Nuki joint includes a hardwood wedge driven in to improve
the tightness of fit between beam and column elements; this additional component might offer
a more normalized tightness of fit across prototypes.
4.5 Conclusions
The rotational stiffness of the Nuki joint was determined using analytical, experimental, and
numerical methods. Two types of materials (Douglas fir and glulam) and two types of fabrication
methods (hand carpentry and digital fabrication) were used for experimental prototypes.
The moment-rotational-displacement plots for all models demonstrated bilinear behavior; an
initial elastic region was observed before wood crushing perpendicular to the grain caused the
joint to enter a plastic region of reduced stiffness. The analytical model tended to under-predict
elastic stiffness values while over-predicting plastic stiffness values and yield moments. The
numerical model for the glulam prototypes predicted absolute stiffness values within 30% of
experimental values. The expected relationship between beam width and rotational stiffness as
predicted by the analytical model was not ideally observed in the experimental prototypes; the
Douglas fir prototypes did not exhibit as much increase in rotational stiffness from beam width
as expected, while the glulam prototypes exhibited a larger increase than expected. The
numerical simulations were more successful at capturing the expected relationship between
beam width and rotational stiffness.
While the overall bilinear behavior may emerge as the only consistent match among the three
models, it is important to note the difficulty of simultaneously tuning all three models to each
other while also acknowledging the limitations and inconsistencies faced in experimental testing.
Given the unpredictable behavior that resulted from the small sample size of experimental
prototypes, the sensitivity to material properties, wood imprefections, and fabrication tolerances
was a clear takeaway from the comparisons between models. The ability to adopt large-scale
statistical models, such as probability distributions of material imperfections, would be important
before the rotational stiffness could be implemented into building codes. As a result, it is
recommended that future work expands the scope of testing for more statistical certainty and
consistency given the variations inherent in both the timber material and the fabrication
methods.
To demonstrate the possible benefits of being able to design with the rotational stiffness
provided by the Nuki joint, the following chapters outline example procedures of designing with
the Nuki joint for gravity and lateral systems. The analytical model developed in this chapter is
used for design.

78
5 Nuki joint in modern timber construction
In the previous chapter, the rotational stiffness of the Nuki joint was characterized using three
methods. This chapter uses the analytical model from the previous chapter to characterize the
rotational stiffness of Nuki joints on beams with varying dimensions, enabling the Nuki joint to
be used in structural design of gravity systems.
A gravity frame design problem is presented for varying spans. Gravity frame designs for 3
different systems are compared: 1) steel beam, 2) glulam beam supported by beam hangers,
and 3) glulam beam supported by Nuki joints.
5.1 Modern mass timber connections
A variety of connection details are available in the present-day mass timber market. For example,
the NDS 2018, the structural code for timber in the United States, lists dowel-type fasteners, split
ring and shear plate connectors, and timber rivets as the three main types of mechanical
connections (American Wood Council 2017).
 Demi Fang 5 Nuki joint in modern timber construction
SMBT 2020

In modern mass timber construction, larger scale buildings sometimes demand more custom
variations of these connections. Usually, these connections use metallic connection components
to join timber elements (Figure 1, Figure 2, Figure 3, Figure 4, Figure 5).
Rather than design a mechanical connection from scratch, in this chapter a pre-engineered beam
hanger is used to represent an example conventional metallic connection for timber systems.
Several pre-engineered beam hanger system products are available in the market. These
products are typically considered nominally pinned and are designed using shear capacity. For
this particular study, the beam hanger catalog from Mass Timber Connections (MTC) is used as
a reference for selecting beam hanger products (MTC Solutions 2020).
5.2 Structural assumptions
A single beam spanning between two columns is considered. The beam section and its
connections at each end to the supporting column are designed. Table 17 summarizes the
constraints assumed.
Table 17. Assumptions for comparative study.

Modulus of elasticity parallel to grain 𝐸𝐸0 1,960,000 psi

Glulam Maximum flexural stress 𝜎𝜎𝑚𝑚𝑚𝑚𝑚𝑚,𝑚𝑚𝑚𝑚𝑚𝑚 4400 psi

properties* Maximum shear stress 𝜏𝜏𝑚𝑚𝑚𝑚𝑚𝑚,𝑚𝑚𝑚𝑚𝑚𝑚 320 psi
Yield strain 𝜀𝜀𝑦𝑦 0.017
10 ft,
15 ft,
Spans 𝐿𝐿 (center-to-center)
20 ft,

Loading 25 ft
conditions Tributary width 𝑏𝑏𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡 𝐿𝐿/2
Dead load 𝐷𝐷 20 psf
Live load 𝐿𝐿 40 psf
Self-weight neglected.
Dimensions Column size 20 in. x 20 in.
*Modulus of elasticity parallel to grain value set to match the value used for glulam prototypes
in Section 4. Maximum flexural and shear stress values are rounded from Nordic Lam Beams
and Headers product catalog (Nordic Engineered Wood 2013).

To calculate the demand on beams supporting the area loads for both service and strength
checks, the linear loads 𝑞𝑞𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢 and 𝑞𝑞𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓 are converted to linear loads accordingly:
𝑞𝑞𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢 = 𝑏𝑏𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡 (𝐷𝐷 + 𝐿𝐿) (54)
𝑞𝑞𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓 = 𝑏𝑏𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡 (1.2𝐷𝐷 + 1.6𝐿𝐿) (55)

80
 Demi Fang 5 Nuki joint in modern timber construction
SMBT 2020

5.3 Beam design

The comparative results for frames designed with beam hangers and frames designed with Nuki
joints are summarized in Figure 46. As will be discussed in greater detail, the rotational restraint
provided by the Nuki joints enable lower midspan moment and subsequently smaller material
quantities. Details on the design method for each are provided in the following subsections.

81
Demi Fang 5 Nuki joint in modern timber construction
SMBT 2020

82
Figure 46. Final beam designs for varying spans using different connection types.
 Demi Fang 5 Nuki joint in modern timber construction
SMBT 2020

5.3.1 Steel beam design

As a benchmark comparison, steel beam sections for each span were designed with shear
connections (nominally pinned) using the Cross Section Optimizer in Karamba3D (Presinger and
Bollinger+Grohmann 2018). The sizes are provided in Table 18.
Table 18. Steel profiles selected for each span.

Span 𝑳𝑳 Steel profile Area

10 ft W4x13 3.83 in.2
15 ft W6x20 5.87 in.2
20 ft W8x40 11.7 in.2
25 ft W10x68 20.0 in.2

5.3.2 Beam design with beam hangers

The beam hanger is nominally pinned on each end. The limit states for beam design are set out
in Table 19.
Table 19. Limit states for the design of a timber beam supported at both ends with a connection of shear
capacity Vallow.

Limit state Condition

1 Maximum flexural stress 𝜎𝜎𝑚𝑚𝑚𝑚𝑚𝑚,𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 < 𝜎𝜎𝑚𝑚𝑚𝑚𝑚𝑚,𝑚𝑚𝑚𝑚𝑚𝑚
2 Maximum shear stress 𝜏𝜏𝑚𝑚𝑚𝑚𝑚𝑚,𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 < 𝜏𝜏𝑚𝑚𝑚𝑚𝑚𝑚,𝑚𝑚𝑚𝑚𝑚𝑚
3 Maximum deflection 𝐿𝐿
𝑤𝑤𝑚𝑚𝑚𝑚𝑚𝑚 <
360
4 Joint shear reaction 𝑉𝑉𝑚𝑚𝑚𝑚𝑚𝑚 < 𝑉𝑉𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎

𝜎𝜎𝑚𝑚𝑚𝑚𝑚𝑚,𝑚𝑚𝑚𝑚𝑚𝑚 and 𝜏𝜏𝑚𝑚𝑚𝑚𝑚𝑚,𝑚𝑚𝑚𝑚𝑚𝑚 are glulam material properties listed in Table 17, 𝐿𝐿 is the span, and 𝑉𝑉𝑎𝑎𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙
is a property of the beam hanger product. All other values are determined through beam analysis
as follows.
Due to the loading conditions, maximum shear is known to occur at the supports. Thus, for limit
state 4, we have
𝑞𝑞𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓 𝐿𝐿
𝑉𝑉𝑚𝑚𝑚𝑚𝑚𝑚 = 𝑉𝑉(0) = 𝑉𝑉(𝐿𝐿) = (56)
2

Due to the loading conditions, the maximum bending moment is known to occur at midspan:
𝐿𝐿 𝑞𝑞𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓 𝐿𝐿2
𝑀𝑀𝑚𝑚𝑚𝑚𝑚𝑚 = 𝑀𝑀 � � = (57)
2 8

84
 Demi Fang 5 Nuki joint in modern timber construction
SMBT 2020

The expressions for 𝑉𝑉𝑚𝑚𝑚𝑚𝑚𝑚 and 𝑀𝑀𝑚𝑚𝑚𝑚𝑚𝑚 can be used to determine 𝜎𝜎𝑚𝑚𝑚𝑚𝑚𝑚,𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 and 𝜏𝜏𝑚𝑚𝑚𝑚𝑚𝑚,𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 to check
the limit states 1 and 2.
ℎ 1
For limit state 1: substituting 𝑀𝑀𝑚𝑚𝑚𝑚𝑚𝑚 from equation (57), 𝑦𝑦 = , and 𝐼𝐼 = 𝑏𝑏ℎ3 (from assuming a
2 12
𝑀𝑀𝑀𝑀
rectangular cross-section) into 𝜎𝜎 = , we have
𝐼𝐼

𝑞𝑞𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓 𝐿𝐿2 ℎ
8 × 2 3𝑞𝑞𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓 𝐿𝐿2
𝜎𝜎𝑚𝑚𝑚𝑚𝑚𝑚,𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 = = (58)
1 3 4𝑏𝑏ℎ2
12 𝑏𝑏ℎ

𝑞𝑞𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓 𝐿𝐿 𝑏𝑏ℎ ℎ 1 𝑉𝑉𝑉𝑉

For limit state 2: substituting 𝑉𝑉 = 𝑉𝑉𝑚𝑚𝑚𝑚𝑚𝑚 = 2
, 𝑄𝑄 = ∗ , 𝐼𝐼 = 𝑏𝑏ℎ3 , and 𝑡𝑡 = 𝑏𝑏 into 𝜏𝜏 = ,
2 4 12 𝐼𝐼𝐼𝐼
we have
𝑞𝑞𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓 𝐿𝐿 𝑏𝑏ℎ2 1 3𝑞𝑞𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓 𝐿𝐿
𝜏𝜏𝑚𝑚𝑚𝑚𝑚𝑚,𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 = × × = (59)
2 8 1 2 3 4𝑏𝑏ℎ
12 𝑏𝑏 ℎ

Finally, for limit state 3: due to the loading conditions, maximum displacement is known to occur
at midspan:
𝐿𝐿 5𝑞𝑞𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢 𝐿𝐿4
𝑤𝑤𝑚𝑚𝑚𝑚𝑚𝑚 = 𝑤𝑤 � � = (60)
2 384𝐸𝐸0 𝐼𝐼

To design the beam, first a beam hanger product was selected for the beam based on the
magnitude of shear reaction. The shear reactions and selected beam hanger products
determined for each span and associated load are summarized in Table 20.
Table 20. Beam hangers selected for each span based on shear demand. Beam hanger product design
values provided in MTC Solutions catalog (MTC Solutions 2020), and S-P-F glulam was assumed.

Shear
Span Shear MTC beam
Fasteners per connector capacity
𝑳𝑳 demand 𝑽𝑽𝒎𝒎𝒎𝒎𝒎𝒎 hanger product
𝑽𝑽𝒂𝒂𝒂𝒂𝒂𝒂𝒂𝒂𝒂𝒂
7 (primary member) + 7
10 ft 2.20 k Ricon S VS 140x60 2.44 k
(secondary member)
16 (primary member) + 16
15 ft 4.95 k Ricon S VS 200x80 6.89 k
(secondary member)
20 ft 8.80 k Megant 310x100 34 9.28 k
30 (primary member) + 30
25 ft 13.8 k Ricon XL 390x80 15.5 k
(secondary member)

85
 Demi Fang 5 Nuki joint in modern timber construction
SMBT 2020

Each beam hanger product has a minimum beam section associated with it. This minimum
section is initially selected as the beam dimensions and the other limit states checked (Table 21).
If needed, the section size is increased slightly to meet the other limit state criteria.
Table 21. Selected beam sections for each span based on beam hanger selection.

Span 𝑳𝑳 Beam section width 𝒃𝒃 Beam section depth 𝒉𝒉

10 ft 4 in. 7 in.
15 ft 4-3/4 in. 11 in.*
20 ft 5-5/8 in. 15-3/4 in.
25 ft 5-5/8 in.** 20-1/2 in.**
*Increased from Ricon S VS 200x60 ℎ𝑚𝑚𝑚𝑚𝑚𝑚 of 9-1/2 in. to meet other limit state criteria.
**Increased from Ricon XL 390x80 𝑏𝑏𝑚𝑚𝑚𝑚𝑚𝑚 of 4-3/4 in. and ℎ𝑚𝑚𝑚𝑚𝑚𝑚 of 17 in. to meet other limit state
criteria.

Key values for the design criteria of each span from the cross-section selections are provided in
Table 22. Moment, displacement, and utilization are also diagrammed in Figure 46(a).
Table 22. Key values for beam design of varying spans supported by beam hangers.

Span 𝑳𝑳
Limit state Value
10 ft 15 ft 20 ft 25 ft
𝜎𝜎𝑚𝑚𝑚𝑚𝑚𝑚,𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 2020 psi 2325 psi 2270 psi 2618 psi
1 Maximum flexural stress 𝜎𝜎𝑚𝑚𝑚𝑚𝑚𝑚,𝑚𝑚𝑚𝑚𝑚𝑚 4400 psi 4400 psi 4400 psi 4400 psi
Utilization 46 % 53 % 52 % 59 %
𝜏𝜏𝑚𝑚𝑚𝑚𝑚𝑚,𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 118 psi 142 psi 149 psi 179 psi
2 Maximum shear stress 𝜏𝜏𝑚𝑚𝑚𝑚𝑚𝑚,𝑚𝑚𝑚𝑚𝑚𝑚 320 psi 320 psi 320 psi 320 psi
Utilization 37 % 44 % 47 % 56 %
𝑤𝑤𝑚𝑚𝑚𝑚𝑚𝑚 0.30 in. 0.49 in. 0.60 in. 0.83 in.
𝐿𝐿
3 Maximum deflection 0.33 in. 0.50 in. 0.67 in. 0.83 in.
360
Utilization 90 % 99 % 90 % 100 %
𝑉𝑉𝑚𝑚𝑚𝑚𝑚𝑚 2.20 k 4.95 k 8.80 k 13.8 k
4 Joint shear reaction 𝑉𝑉𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 2.44 k 5.89 k 9.28 k 15.5 k
Utilization 90 % 72 % 95 % 89 %

86
 Demi Fang 5 Nuki joint in modern timber construction
SMBT 2020

In all spans, either deflection or joint shear reaction (limit states 3 or 4) govern the design of the
beam section. For 15 ft and 25 ft spans, the beam section had to be increased from the minimum
required by the selected beam hanger in order to meet the deflection limit. The 15 ft span has
a relatively low utilization because it happens to have a shear demand that falls in an intermediate
range between shear capacities of beam hanger products for S-P-F glulam.
5.3.3 Beam design with Nuki joints
The rotational stiffness of the Nuki joint depends on the dimensions of the timber elements, so
the timber beam must be designed in parallel with the rotational stiffness. In this chapter, the
rotational stiffness of the Nuki joint in the elastic region is called 𝑘𝑘, and the width and depth of
the beam are called 𝑏𝑏 and ℎ, respectively. A summary of the limit states for beam design are
provided in Table 23.
Table 23. Limit states for the design of a timber beam supported at both ends with a Nuki joint of known
rotational stiffness k.

Limit state Condition

1 Maximum flexural stress 𝜎𝜎𝑚𝑚𝑚𝑚𝑚𝑚,𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 < 𝜎𝜎𝑚𝑚𝑚𝑚𝑚𝑚,𝑚𝑚𝑚𝑚𝑚𝑚
2 Maximum shear stress 𝜏𝜏𝑚𝑚𝑚𝑚𝑚𝑚,𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 < 𝜏𝜏𝑚𝑚𝑚𝑚𝑚𝑚,𝑚𝑚𝑚𝑚𝑚𝑚
3 Maximum deflection 𝐿𝐿
𝑤𝑤𝑚𝑚𝑚𝑚𝑚𝑚 <
360
4 Joint moment reaction 𝑀𝑀𝑅𝑅 < 𝑀𝑀𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎
where 𝑀𝑀𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 = 𝑘𝑘𝜃𝜃𝑦𝑦
𝜀𝜀𝑦𝑦 ℎ
and 𝜃𝜃𝑦𝑦 = tan−1 �𝐶𝐶 �
𝑑𝑑 /2

𝜎𝜎𝑚𝑚𝑚𝑚𝑚𝑚,𝑚𝑚𝑚𝑚𝑚𝑚 and 𝜏𝜏𝑚𝑚𝑚𝑚𝑚𝑚,𝑚𝑚𝑚𝑚𝑚𝑚 are glulam material properties listed in Table 17.
𝑘𝑘 is determined by the analytical model of the Nuki joint presented in Section 3.2. The model as
originally presented assumes some “ended embedment length” representing the length past
which the beam pierces through the far face of the column. For this case study, it is assumed that
the beam stops at the far face of the column, meaning an “ended embedment length” of 0. All
terms related to ended embedment (subscript 𝑒𝑒) presented in Section 3.2 are taken as 0.
Because of the iterative nature of the design process, it was more convenient to generate a
reference to look up rotational stiffnesses associated with different section sizes. These values
were generated using a variation of the script presented in Appendix 1 (where ended
embedment terms are eliminated and beam dimensions 𝑏𝑏 and ℎ are input for 𝐵𝐵𝑤𝑤 and 𝐵𝐵𝑑𝑑 ,
respectively) and is provided in Figure 47. The values are valid for a column depth 𝐶𝐶𝑑𝑑 of 20 inches.
(𝐶𝐶𝑑𝑑 is the column dimension parallel to the main axis of the beam; see Figure 18).

87
 Demi Fang 5 Nuki joint in modern timber construction
SMBT 2020

Figure 47. Graphic to assist with Nuki beam and joint design, charting the rotational stiffness of Nuki
joints of varying widths b and depths h in the range 1 < h/b < 4. These values are calculated for a
column depth (dimension Cd, based on terminology of Section 3) of 20 inches.

To design the beam section, an initial size is selected, and 𝑘𝑘 is determined by looking up the
value in Figure 47. Unlike the beam hanger design process in Section 5.3.2, the beam supported
by Nuki joints cannot be designed using the governing design values for a simply supported
beam. With a known elastic rotational stiffness 𝑘𝑘 supporting the beam and the loading conditions
set out in Table 17, the moment reactions and midspan deflection of the beam supported on

88
 Demi Fang 5 Nuki joint in modern timber construction
SMBT 2020

both ends with a Nuki joint can be derived; details of these calculations are set out in Appendix
3. For ease of reference, the expressions for moment reaction magnitude 𝑀𝑀𝑅𝑅 and midspan
𝐿𝐿
deflection 𝑤𝑤 � � are repeated here, with a substitution of 𝐸𝐸 = 𝐸𝐸0 to indicate the stiffness parallel
2
to grain as the relevant bending stiffness in glulam:
𝑞𝑞𝐿𝐿3
𝑀𝑀𝑅𝑅 = (61)
𝐸𝐸 𝐼𝐼 𝐿𝐿
24 � 0 + 2�
𝑘𝑘
𝐿𝐿 𝐿𝐿 𝑀𝑀𝑅𝑅 𝑀𝑀𝑅𝑅 𝐿𝐿 𝑞𝑞𝐿𝐿3
𝑤𝑤 � � = �−𝐸𝐸0 𝐼𝐼 − + � (62)
2 𝐸𝐸0 𝐼𝐼 2𝑘𝑘 8 128
𝐿𝐿
Since it is known that 𝑤𝑤𝑚𝑚𝑚𝑚𝑚𝑚 = 𝑤𝑤 � � under these loading conditions, these expressions can be
2
directly used to evaluate limit states 3 and 4 from Table 23.
Due to the loading conditions, maximum shear is known to occur at the supports with 𝑉𝑉(0) =
𝑞𝑞𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓 𝐿𝐿
𝑉𝑉(𝐿𝐿) = , and the maximum bending moment is known to occur either at the supports
2
𝐿𝐿
or the midpoint: 𝑀𝑀𝑚𝑚𝑚𝑚𝑚𝑚 = max �𝑀𝑀𝑅𝑅 , 𝑀𝑀 � ��.
2

These expressions for 𝑉𝑉𝑚𝑚𝑚𝑚𝑚𝑚 and 𝑀𝑀𝑚𝑚𝑚𝑚𝑚𝑚 can be used to determine 𝜎𝜎𝑚𝑚𝑚𝑚𝑚𝑚,𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 and 𝜏𝜏𝑚𝑚𝑚𝑚𝑚𝑚,𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 to
check the limit states 1 and 2 from Table 23.
𝐿𝐿 ℎ 1
For limit state 1: substituting 𝑀𝑀 = 𝑀𝑀𝑚𝑚𝑚𝑚𝑚𝑚 = max �𝑀𝑀𝑅𝑅 , 𝑀𝑀 � ��, 𝑦𝑦 = , and 𝐼𝐼 = 𝑏𝑏ℎ3 assuming a
2 2 12
𝑀𝑀𝑀𝑀
rectangular cross-section into 𝜎𝜎 = , we have
𝐼𝐼

𝐿𝐿 ℎ 𝐿𝐿
max �𝑀𝑀𝑅𝑅 , 𝑀𝑀 � �� × max �𝑀𝑀𝑅𝑅 , 𝑀𝑀 � �� × 6
2 2 2 (63)
𝜎𝜎𝑚𝑚𝑚𝑚𝑚𝑚,𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 = =
1 𝑏𝑏ℎ2
𝑏𝑏ℎ3
12

𝑞𝑞𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓 𝐿𝐿 𝑏𝑏ℎ ℎ 1 𝑉𝑉𝑉𝑉

For limit state 2: substituting 𝑉𝑉 = 2
, 𝑄𝑄 = ∗ , 𝐼𝐼 = 𝑏𝑏ℎ3 , and 𝑡𝑡 = 𝑏𝑏 into 𝜏𝜏 = , we have
2 4 12 𝐼𝐼𝐼𝐼

𝑞𝑞𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓 𝐿𝐿 𝑏𝑏ℎ2 1 3𝑞𝑞𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓 𝐿𝐿

𝜏𝜏𝑚𝑚𝑚𝑚𝑚𝑚,𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 = × × = (64)
2 8 1 2 3 4𝑏𝑏ℎ
12 𝑏𝑏 ℎ
For a given span, a range of cross-sections (i.e. combinations of 𝑏𝑏 and ℎ) can be determined.
This set of cross-sections can be generated by finding the minimum ℎ for several values of 𝑏𝑏 that
satisfies all limit states. A set of cross-sections that meet all limit states for each span, within the
range 1 ≤ ℎ/𝑏𝑏 < 4 and at the increments provided in Figure 47, are listed in Table 24.

89
 Demi Fang 5 Nuki joint in modern timber construction
SMBT 2020

Table 24. Beam cross sections satisfying all limit states when supported by Nuki joints. Cross sections are
selected by picking a width b and finding the minimum depth h that satisfies all limit states, within the
range 1 < h/b < 4.

Span Beam section Beam section Beam section area Beam section ratio
𝑳𝑳 width 𝒃𝒃 depth 𝒉𝒉 𝒃𝒃𝒃𝒃 𝒉𝒉/𝒃𝒃
2 in. 5-1/2 in. 11.0 in.2 2.75
10 ft
3 in. 4-1/2 in. 13.5 in.2 1.50
3 in. 9-1/2 in. 28.5 in.2 3.17
15 ft 3-1/2 in. 8-1/2 in. 29.8 in.2 2.43
4-1/2 in. 7-1/2 in. 33.8 in.2 1.67
4 in. 14-1/2 in. 58.0 in.2 3.63
20 ft 4-1/2 in. 13-1/2 in. 60.8 in.2 3.00
5-1/2 in. 12-1/2 in. 68.8 in.2 2.27
25 ft 5 in. 19-1/2 in. 97.5 in.2 3.9

The data displayed in Table 24 make it clear that for each span, the most materially efficient
sections (lowest section area 𝑏𝑏ℎ) are the narrow and tall sections (high ℎ/𝑏𝑏). However, there is a
practical limit on how narrow a section can be due to lateral torsional buckling concerns, which
is why a maximum ℎ/𝑏𝑏 of 4 was set using engineering judgment when selecting cross-sections
for each span.
Table 25 lists the cross-section dimensions ultimately selected for each span supported by Nuki
joints, along with the section’s rotational stiffness.
Table 25. Selected beam sections for each span based on limit states for beams supported by Nuki
joints, and associated rotational stiffness in the elastic region.

Span 𝑳𝑳 Beam section width 𝒃𝒃 Beam section depth 𝒉𝒉 Rotational stiffness 𝒌𝒌

10 ft 2 in. 5-1/2 in. 1376 k-ft/rad
15 ft 3 in. 9-1/2 in. 1365 k-ft/rad
20 ft 4 in. 14-1/2 in. 1392 k-ft/rad
25 ft 5 in. 19-1/2 in. 1461 k-ft/rad

Key values for the design criteria of each span from the cross-section selections are provided in
Table 26. Moment, displacement, and utilization are also diagrammed in Figure 46(b).

90
 Demi Fang 5 Nuki joint in modern timber construction
SMBT 2020

Table 26. Key values for beam design of varying spans supported by Nuki joints.

Span 𝑳𝑳
Limit state Value
10 ft 15 ft 20 ft 25 ft
𝜎𝜎𝑚𝑚𝑚𝑚𝑚𝑚,𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 4137 psi 2561 psi 2507 psi 2598 psi
1 Maximum flexural stress 𝜎𝜎𝑚𝑚𝑚𝑚𝑚𝑚,𝑚𝑚𝑚𝑚𝑚𝑚 4400 psi 4400 psi 4400 psi 4400 psi
Utilization 94 % 58 % 57 % 59 %
𝜏𝜏𝑚𝑚𝑚𝑚𝑚𝑚,𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 300 psi 261 psi 228 psi 212 psi
2 Maximum shear stress 𝜏𝜏𝑚𝑚𝑚𝑚𝑚𝑚,𝑚𝑚𝑚𝑚𝑚𝑚 320 psi 320 psi 320 psi 320 psi
Utilization 94 % 81 % 71 % 66 %
𝑤𝑤𝑚𝑚𝑚𝑚𝑚𝑚 0.30 in. 0.46 in. 0.65 in. 0.82 in.
𝐿𝐿
3 Maximum deflection 0.33 in. 0.50 in. 0.67 in. 0.83 in.
360
Utilization 90 % 92 % 97 % 99 %
𝑀𝑀𝑅𝑅 3.48 k-ft 9.63 k-ft 14.7 k-ft 17.4 k-ft
4 Joint moment reaction 𝑀𝑀𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 12.9 k-ft 22.0 k-ft 34.3 k-ft 48.4 k-ft
Utilization 27 % 44 % 43 % 36 %

For all spans above 10 ft, deflection (limit state 3) definitively governs the design of the beam
section. For the 10 ft span, maximum flexural stress and maximum shear stress (limit states 1 and
2) govern slightly more than deflection.
5.4 Comparing embodied carbon for all gravity frame design methods
In the previous section, varying design methods were used to design a gravity frame for the same
span and loading conditions. This section quantifies and compares the embodied carbon of each
system.
5.4.1 Assumptions
To calculate the associated embodied carbon for each system, material densities and embodied
carbon coefficients must be assumed. The assumed values for respective frame components are
provided in Table 27.

91
 Demi Fang 5 Nuki joint in modern timber construction
SMBT 2020

Table 27. Material assumptions for calculation of embodied carbon. Embodied carbon coefficients were
obtained from Jones and Hammond 2019.

Component Material Density Embodied carbon

coefficient
Steel beam Steel, section (world 0.132 kg/in.3 1.55 kg CO2e / kg
profiles average)
Beam hanger Steel, wire rod (world 0.132 kg/in.3 2.27 kg CO2e / kg
fasteners average)
RICON S VS beam Steel, plate (world average) 0.132 kg/in.3 2.46 kg CO2e / kg
hangers
MEGANT beam Aluminum (world average) 0.0442 kg/in.3 13.1 kg CO2e / kg
hangers
Beams Timber, glulam (no carbon 0.00688 kg/in.3 0.512 kg CO2e / kg
storage)

The same column size was used for each gravity frame, so column volumes were omitted. In the
Nuki joint systems, the columns are mortised to insert the beams; the volumes of timber beam
intersecting the column were excluded from volume calculations.
For the steel beam profiles, volumes of shear connectors were omitted because the amount of
additional metallic material was negligible relative to the volume of steel section with a
comparable EC coefficient.
For the beam hangers, volumes of hex nuts and washers were omitted.
Volumes for the beam hanger products were not provided by the manufacturer, but sufficient
dimensions were presented in the product design guide to estimate volume quantities. Details
on these volume estimates of beam hanger products are provided in Appendix 4.
5.4.2 Quantities
Volume and embodied carbon quantities for the 10 ft, 15 ft, 20 ft, and 25 ft spans are presented
in Table 28, Table 29, Table 30, and Table 31, respectively. Embodied carbon quantities are
shown with perspective renders of each frame design in Figure 48.

92
 Table 28. Quantities of different gravity frame designs for a 10 ft span.

10 ft span Steel Beam hanger Nuki, full section

Beam section W6x20 4 in. x 7 in. 2 in. x 5-1/2 in.
Timber, glulam area 0 in.2 28 in.2 11 in.2
Steel, section area 3.83 in.2 0 in.2 0 in.2
Steel, section 459.6 in.3 0 in.3 0 in.3
Steel, wire rod 0 in.3 10.5 in.3 0 in.3
Steel, plate 0 in.3 9.8 in.3 0 in.3
Aluminum 0 in.3 0 in.3 0 in.3
Timber, glulam 0 in.3 3360 in.3 1320 in.3
Steel, section 94.0 kg CO2e 0 kg CO2e 0 kg CO2e
Steel, wire rod 0 kg CO2e 2.65 kg CO2e 0 kg CO2e
Steel, plate 0 kg CO2e 3.18 kg CO2e 0 kg CO2e
Aluminum 0 kg CO2e 0 kg CO2e 0 kg CO2e
Timber, glulam 0 kg CO2e 11.8 kg CO2e 4.65 kg CO2e
TOTAL EC 94.0 kg CO2e 18.2 kg CO2e 4.65 kg CO2e
Demi Fang 5 Nuki joint in modern timber construction
SMBT 2020

Table 29. Quantities of different gravity frame designs for a 15 ft span.

15 ft span Steel Beam hanger Nuki, full section

4-3/4 in. x
Beam section W4x13 3 in. x 9-1/2 in.
11 in.
Timber, glulam area 0 in.2 52.25 in.2 28.5 in.2
Steel, section area 5.87 in.2 0 in.2 0 in.2
Steel, section 1057 in.3 0 in.3 0 in.3
Steel, wire rod 0 in.3 46.4 in.3 0 in.3
Steel, plate 0 in.3 18.5 in.3 0 in.3
Aluminum 0 in.3 0 in.3 0 in.3
Timber, glulam 0 in.3 9405 in.3 5130 in.3
Steel, section 216.2 kg CO2e 0 kg CO2e 0 kg CO2e
Steel, wire rod 0 kg CO2e 13.9 kg CO2e 0 kg CO2e
Steel, plate 0 kg CO2e 5.99 kg CO2e 0 kg CO2e
Aluminum 0 kg CO2e 0 kg CO2e 0 kg CO2e
Timber, glulam 0 kg CO2e 33.1 kg CO2e 18.1 kg CO2e
TOTAL EC 216.2 kg CO2e 53.0 kg CO2e 18.1 kg CO2e

94
Demi Fang 5 Nuki joint in modern timber construction
SMBT 2020

Table 30. Quantities of different gravity frame designs for a 20 ft span.

20 ft span Steel Beam hanger Nuki, full section

5-5/8 in. x
Beam section W8x40 4 in. x 14-1/2 in.
15-3/4 in.
Timber, glulam area 0 in.2 88.6 in.2 58 in.2
Steel, section area 11.7 in.2 0 in.2 0 in.2
Steel, section 2808 in.3 0 in.3 0 in.3
Steel, wire rod 0 in.3 57.3 in.3 0 in.3
Steel, plate 0 in.3 0 in.3 0 in.3
Aluminum 0 in.3 161.0 in.3 0 in.3
Timber, glulam 0 in.3 21,263 in.3 13,920 in.3
Steel, section 574.5 kg CO2e 0 kg CO2e 0 kg CO2e
Steel, wire rod 0 kg CO2e 17.2 kg CO2e 0 kg CO2e
Steel, plate 0 kg CO2e 0 kg CO2e 0 kg CO2e
Aluminum 0 kg CO2e 93.2 kg CO2e 0 kg CO2e
Timber, glulam 0 kg CO2e 74.9 kg CO2e 49.0 kg CO2e
TOTAL EC 574.5 kg CO2e 185.3 kg CO2e 49.0 kg CO2e

95
Demi Fang 5 Nuki joint in modern timber construction
SMBT 2020

Table 31. Quantities of different gravity frame designs for a 25 ft span.

25 ft span Steel Beam hanger Nuki, full section

5-5/8 in. x
Beam section W10x68 5 in. x 19-1/2 in.
20-1/2 in.
Timber, glulam area 0 in.2 115.3 in.2 97.5 in.2
Steel, section area 20.0 in.2 0 in.2 0 in.2
Steel, section 6000 in.3 0 in.3 0 in.3
Steel, wire rod 0 in.3 87.0 in.3 0 in.3
Steel, plate 0 in.3 56.3 in.3 0 in.3
Aluminum 0 in.3 0 in.3 0 in.3
Timber, glulam 0 in.3 34,594 in.3 29,250 in.3
Steel, section 1228 kg CO2e 0 kg CO2e 0 kg CO2e
Steel, wire rod 0 kg CO2e 26.1 kg CO2e 0 kg CO2e
Steel, plate 0 kg CO2e 18.3 kg CO2e 0 kg CO2e
Aluminum 0 kg CO2e 0 kg CO2e 0 kg CO2e
Timber, glulam 0 kg CO2e 121.9 kg CO2e 103.0 kg CO2e
TOTAL EC 1228 kg CO2e 166.2 kg CO2e 103.0 kg CO2e

96
Figure 48. Embodied carbon comparison between gravity frame design methods. Embodied carbon scales for steel sections are abbreviated in order to keep scales practical for other methods. For beam hangers, only fasteners into the
primary member are illustrated; fasteners into the secondary member are still included in quantities calculations and are reflected in the displayed embodied carbon quantities.
 Demi Fang 5 Nuki joint in modern timber construction
SMBT 2020

The embodied carbon values are also plotted against span in Figure 49.

Figure 49. Embodied carbon against span for different design methods. The open dot acknowledges
that a beam hanger product from a different line, the MEGANT series with more aluminum than the
Ricon series, was used for the 20 ft span.

To aid with discussion, ratios between embodied carbon values from different design methods
are calculated and provided in Table 32.
Table 32. Ratios between embodied carbon values for different spans and methods.

EC MTC / EC Nuki /
Span [ft]
EC Steel EC Beam hangers

10 0.19 0.26
15 0.25 0.34
20 0.32 0.26
25 0.14 0.62

5.5 Discussion: comparing quantities across frame designs

The use of timber for the beam element is indisputably an advantage from an embodied carbon
perspective; comparing the steel beams to the timber beams supported by beam hangers,
embodied carbon values were improved by as much as 86% in the 25 ft span scenario (Table 32).
Furthermore, the rate of increase in embodied carbon when designing with steel sections
appears to be much higher than for either timber system (Figure 49).
98
 Demi Fang 5 Nuki joint in modern timber construction
SMBT 2020

The rotational stiffness provided by the Nuki joints resulted in a reduced midspan moment
(Figure 46). Given most of the beams were governed by deflection limits, a reduced midspan
moment meant that the beams designed with the Nuki joints could have smaller cross-sections.
Between the two connection types, the Nuki joint enabled volume reduction in timber elements
through joint stiffness alone. Furthermore, the Nuki joint involves no metallic parts, meaning that
it could further outcompete the beam hangers in embodied carbon quantities. These benefits
added up to 38% to 74% improvement in embodied carbon quantities (Table 32).
A few assumptions were made in this comparative study that are acknowledged to influence the
results. Reflections on the effects of those assumptions are provided below:
• 𝑀𝑀𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 of each beam section depended on the 𝜃𝜃𝑦𝑦 value associated with each beam
section. The selection of 𝜃𝜃𝑦𝑦 was based on the hard limitation that it would not be
tolerable for the joint to enter the plastic region. There was also no safety factor applied
to the rotational stiffness 𝑘𝑘 in this study (i.e. 𝑀𝑀𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 = 1.0𝑀𝑀𝑦𝑦 ). More experimental testing
to calibrate the models in Section 4 are needed to determine an appropriate safety factor.
The analytical model used to design the Nuki joints were found in Section 4 to be
conservative in stiffness 𝑘𝑘 but unconservative in yield moment 𝑀𝑀𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 .
• The rotational stiffness 𝑘𝑘 of each beam section provided by the Nuki joint depended on
the column dimension parallel to the beam axis, 𝐶𝐶𝑑𝑑 , which was fixed at 20 inches for all
spans in this study. If 𝐶𝐶𝑑𝑑 was set to vary linearly with span, there might be increasing end
rigidity benefits at increasing spans than observed in these results. However, there is
likely a practical limit on how column size varies with span. Moreover, this relationship
between 𝐶𝐶𝑑𝑑 and 𝑘𝑘 is specific to the Nuki joint; other joinery geometries may not derive
as much stiffness from column dimensions.
• Due to the particular geometry of the Nuki joint, the construction sequence for gravity
frames using the Nuki joint may not be immediately compatible with modern methods of
timber construction. However, the methods may be applied to other joinery connections
whose rotational stiffnesses could be characterized in a similar way.
• Self-weight was neglected in this study. Accounting for the self-weight of the beam in the
clear span would likely increase the section size needed to meet the limit states, but it is
not immediately clear how it might change the comparative results between design
methods.
• The tributary width was set to increase linearly with span, but the ratio of span to tributary
width may depend on the flooring system.
• Embodied carbon values do not necessarily reflect time or labor costs.
• The pre-engineered beam hanger was assumed to represent a conventional fastener
connection. This methodology may be repeated on other types of fastener connections.
• The scope of this LCA is cradle-to-gate.
• The effects of continuous spans and lateral loads have yet to be evaluated in detail. A
preliminary assessment on the use of the Nuki joint in a lateral system with the beam and
joint sizes designed in this section is illustrated in Figure 50. Drift governs and is
dependent on the number of continuous spans and the number of floors. The beam
designed under gravity loads was adequate for the illustrated number of bays and floors
under unfactored gravity loads and wind loads while reaching only about 50% of joint
moment capacity. Larger beam sizes could be used for structures with higher slenderness
ratios to meet drift limits.
99
 Demi Fang 5 Nuki joint in modern timber construction
SMBT 2020

• The comparison does not reflect more qualitative characteristics, such as the architectural
asset of creating more clear space achieved by reduced beam depth, the benefits of a
non-destructive assembly, the visual continuity of a single-material system, and other
opportunities mentioned in Section 1.4.

Figure 50. Preliminary assessment on the use of Nuki joints with rotational restraint for lateral systems,
elevation view. 30 psf wind loads were applied as point loads assuming a 10 ft tributary width, as well as
unfactored gravity loads outlined in Section 5.2. Timber columns measuring 20” x 20” were modeled.

5.6 Conclusions
The sustainability benefits of using timber over conventional carbon-intensive building materials
such as steel as primary elements in structural systems are relatively well established (Section
1.4.3). However, studies seldom examine the embodied carbon implications of using the metallic
fasteners conventionally used in modern timber construction in contrast to all-timber joinery
connections used in more historic timber traditions.
To make these comparisons, gravity frames were designed using different beam and connection
approaches for the same loading conditions across varying single spans. The quantity of
embodied carbon required for each design method was compared.
As discussed in Section 2.4, two main pathways for the design of low-carbon structures are (1)
reducing structural material quantities and (2) designing with materials with lower embodied
100
 Demi Fang 5 Nuki joint in modern timber construction
SMBT 2020

carbon (De Wolf 2017). In line with pathway (2), in this chapter, the benefit of using steel over
timber as a beam element was confirmed. The strategy for introducing Nuki joints followed
pathway (1) in that the added rotational stiffness allowed for a smaller cross-section of timber for
those beams supported by Nuki joints. Combined with the lack of metallic fasteners in the system,
the Nuki joint frames demonstrated decisive improvements in embodied carbon from the beam
hangers – up to 74% improvement in some spans.
The results of this comparative study imply that joinery connections could represent a sustainable
alternative to conventional metallic fastener connections used in modern timber construction.
Moreover, the results indicate that structural connection quantities might represent an important
consideration in evaluating the sustainability of different structural systems. The methodology of
the comparative study can be repeated on other joinery geometries and other fastener
connections, setting a precedent as a workflow for evaluating the sustainability of connections
in structural design.

101
Demi Fang 5 Nuki joint in modern timber construction
SMBT 2020

102
6 Conclusions
6.1 Summary of contributions
This thesis revisits the timber joinery traditions of global history through a lens of modern
structural analysis. Models for the rotational stiffness of the Japanese Nuki joint, a mortise-and-
tenon-type joinery, were developed based on literature (Section 3) and their results compared
on varying beam widths (Section 4). All exhibited the expected bilinear stiffness behavior due to
the unique properties of wood crushed perpendicular to the grain. The accuracy of absolute and
relative stiffness values and yield moment across models was more difficult to verify due to
limitations in experimental testing. In order to proceed with structural design using the Nuki joint,
it was noted that the analytical model tended to under-predict the elastic rotational stiffness of
the joint (𝑘𝑘1 ) and over-predict the moment at which the joint would yield to plastic behavior (𝑀𝑀𝑦𝑦 ).
The analytical model was then used to design a gravity frame under loading across varying spans,
representing a new contribution to literature in evaluating the effect of structural all-timber
connections on sustainability in early-stage structural design (Section 5). The quantities of
 Demi Fang 6 Conclusions
SMBT 2020

material required for each gravity frame was compared to an analogous frame supported by
beam hangers, a conventional structural connection in modern timber construction. The
rotational stiffness provided by the Nuki joint reduces the midspan moment, enabling a smaller
timber section to be selected for the beam supported by Nuki joints than for the beam supported
by nominally pinned beam hangers. Combined with the absence of metallic parts in the Nuki
joint, the embodied carbon savings from using Nuki joints ranged from 38% to 76%.

Figure 51. Summary of findings comparing the embodied carbon of a beam supported by conventional
beam hangers to that of a beam supported by Nuki joints. In the 20 ft span pictured, embodied carbon
savings of 76% are observed.

6.2 Potential impact

This thesis identifies joinery connections in modern timber construction as a potential
opportunity for lower-carbon structures. Barriers to entry into modern timber structural systems
may include the relative inaccessibility of joint characterization to structural designers, as well as
an ambiguity around the benefits of selecting joinery connections over existing conventions in
structural connections for timber elements. By presenting a methodology for characterizing the
behavior of joinery connections and demonstrating the carbon savings achieved by selecting
joinery over conventional connections, this thesis clarifies the incentives and benefits for using
joinery connections in modern timber construction. If adapted at a larger scale, these
applications of joinery connections have the potential to help reduce the carbon footprint of
structural systems at large.

104
 Demi Fang 6 Conclusions
SMBT 2020

6.3 Limitations and future work

The difficulty of calibrating experimental methods with analytical and numerical methods
suggests that much more experimental testing is needed due to variability in the wood (seen in
both raw timber and engineered timber) and fabrication processes (seen in both hand carpentry
and digital fabrication). The statistical analysis and scale of testing needed to recommend safety
factors for design is beyond the scope of this thesis.
More work is also needed to apply these methods to joinery geometries that may be best
adaptable by modern timber fabrication facilities. It is envisioned that a catalog of joinery
connections with associated design values could be valuable for structural designers to apply
these joints at a larger scale.
6.4 Concluding remarks
This thesis makes two major arguments related to timber joinery in modern timber construction:
1) timber joinery connections can be evaluated using modern engineering models for structural
design, and 2) if re-introduced as structural connections, timber joinery can play a significant role
in reducing the embodied carbon of structures. Along with other potential opportunities such as
seismic performance, improved physiology of people inside an all-timber environment, and the
longevity enabled by non-destructive assembly, it is clear that the use of timber joinery for
structural connections are not to be overlooked in the current search for low-carbon strategies
for structural design.

105
Demi Fang 6 Conclusions
SMBT 2020

106
References
American Wood Council. 2017. “National Design Specification (NDS) for Wood Construction
2018 Edition.” American Wood Council.
Architectural Institute of Japan (AIJ). 2010. Fundamental Theory of Timber Engineering. Tokyo,
Japan: Maruzen.
“Architecture 2030.” n.d. Accessed July 12, 2019. [Link]
ASTM International. 2014. “ASTM Standard D143-14: Standard Test Methods for Small Clear
Specimens of Timber.” Standard. West Conshohocken, PA.
Benson, Tedd. 1997. The Timber-Frame Home - Design, Construction, Finishing. 2nd ed.
Taunton Press.
Böhme, Luis Felipe González, Francisco Quitral Zapata, and Sandro Maino Ansaldo. 2017.
“Roboticus Tignarius: Robotic Reproduction of Traditional Timber Joints for the
Reconstruction of the Architectural Heritage of Valparaíso.” Construction Robotics, July,
1–8. [Link]
Brownell, Blaine. 2016. “T3 Becomes the First Modern Tall Wood Building in the U.S.” Architect.
November 8, 2016. [Link]
first-modern-tall-wood-building-in-the-us_o.
Browning, William, Cathering Ryan, and Joseph Clancy. 2014. “14 Patterns of Biophilic Design -
Improving Health and Well-Being in the Built Environment.” Terrapin Bright Green.
 Demi Fang References
SMBT 2020

Brungraber, Robert Lyman. 1985. “Traditional Timber Joinery: A Modern Analysis.” Stanford
University. [Link]
Building Transparency. 2019. “Embodied Carbon in Construction Calculator (EC3).” October 8,
2019. [Link]
Chang, Wen-Shao, and Min-Fu Hsu. 2007. “Rotational Performance of Traditional Nuki Joints
with Gap II: The Behavior of Butted Nuki Joint and Its Comparison with Continuous Nuki
Joint.” Journal of Wood Science 53 (5): 401–7. [Link]
0880-1.
Chang, Wen-Shao, Min-Fu Hsu, and Kohei Komatsu. 2006. “Rotational Performance of
Traditional Nuki Joints with Gap I: Theory and Verification.” Journal of Wood Science 52
(1): 58–62. [Link]
Chang, Wen-Shao, Takehiro Wakita, Akihisa Kitamori, Kohei Komatsu, Yasuo Kataoka, Min-Fu
Hsu, and web-support@[Link]. 2012. “Dynamic Properties of a Traditional Complex
Bracket Sets in an Oriental Temple.” In World Conference on Timber Engineering 2012,
WCTE [Link]. 5., 460–63. University of Bath. [Link]
Chen, Jinyong, Tieying Li, Qingshan Yang, Xiwang Shi, and Yanxia Zhao. 2018. “Degradation
Laws of Hysteretic Behaviour for Historical Timber Buildings Based on Pseudo-Static
Tests.” Engineering Structures 156 (February): 480–89.
[Link]
Chen, Ling-kun, Sheng-cai Li, Yan-ting Wang, You-jun Zhao, Ming Zhang, Xing-yu Song, Xiao-
wei Li, Teng Wu, and Li-zhong Jiang. 2017. “Experimental Study on the Seismic
Behaviour of Mortise–Tenon Joints of the Ancient Timbers.” Structural Engineering
International 27 (4): 512–19. [Link]
Chen Zhiyong, Zhu Enchun, Pan Jinglong, and Wu Guofang. 2016. “Energy-Dissipation
Performance of Typical Beam-Column Joints in Yingxian Wood Pagoda: Experimental
Study.” Journal of Performance of Constructed Facilities 30 (3): 04015028.
[Link]
CORE Studio. 2020. Beacon. [Link]
Court, Brian. 2012. “Timber Framing at the Bullitt Center | Bullitt Center.” March 19, 2012.
[Link]
D’Ayala, Dina F., and Pin Hui Tsai. 2008. “Seismic Vulnerability of Historic Dieh–Dou Timber
Structures in Taiwan.” Engineering Structures, Seismic reliability, analysis, and protection
of historic buildings and heritage sites, 30 (8): 2101–13. [Link]
De Wolf, Catherine Elvire Lieve. 2017. “Low Carbon Pathways for Structural Design : Embodied
Life Cycle Impacts of Building Structures.” Thesis, Massachusetts Institute of Technology.
[Link]
De Wolf, Catherine, Endrit Hoxha, Alexander Hollberg, Corentin Fivet, and John Ochsendorf.
2020. “Database of Embodied Quantity Outputs: Lowering Material Impacts Through
Engineering.” Journal of Architectural Engineering 26 (3): 04020016.
[Link]
Erdil, Y. Z., A. Kasal, and C. A. Eckelman. 2005. “Bending Moment Capacity of Rectangular
Mortise and Tenon Furniture Joints.” Forest Products Journal 55 (12): 209–13.
Erikson, Robert George. 2003. “Behavior of Traditional Timber Frame Structures Subjected to
Lateral Load.” Ph.D., United States -- Wyoming: University of Wyoming.
[Link]
48EEPQ/1.

108
 Demi Fang References
SMBT 2020

Fang, Demi, Julieta Moradei, Jan Brütting, Aliz Fischer, Daniel K Landez, Benshun Shao, Nick
Sherrow-Groves, Corentin Fivet, and Caitlin Mueller. 2019. “Modern Timber Design
Approaches for Traditional Japanese Architecture: Analytical, Experimental, and
Numerical Approaches for the Nuki Joint.” In Proceedings of the International
Association for Shell and Spatial Structures (IASS) Symposium. Barcelona, Spain.
Fang, Demi, Caitlin Mueller, Jan Brütting, Corentin Fivet, and Julieta Moradei. 2019. “Rotational
Stiffness in Timber Joinery Connections: Analytical and Experimental Characterizations of
the Nuki Joint.” In Structures and Architecture: Bridging the Gap and Crossing Borders,
edited by Paulo J.S. Cruz, 229–36. Lisbon, Portugal: CRC Press.
[Link]
Foliente, G. C. 2000. “History of Timber Construction.” ASTM International, January.
[Link]
Forest Products Laboratory. 2010. “Wood Handbook: Wood as an Engineering Material.”
General Technical Report FPL-GTR-190. Madison, WI: U.S. Department of Agriculture,
Forest Service, Forest Products Laboratory.
Fujita, Kaori, Isao Sakamoto, Yoshimitsu Ohashi, and Masahiko Kimura. 2000. “Static and
Dynamic Loading Tests of Bracket Complexes Used in Traditional Timber Structures in
Japan.” In Proceedings of the 12th World Conference on Earthquake Engineering. Vol.
30. Auckland, New Zealand.
Guan, Z. W., A. Kitamori, and K. Komatsu. 2008. “Experimental Study and Finite Element
Modelling of Japanese ‘Nuki’ Joints — Part Two: Racking Resistance Subjected to
Different Wedge Configurations.” Engineering Structures 30 (7): 2041–49.
[Link]
Gustavsson, Leif, Kim Pingoud, and Roger Sathre. 2006. “Carbon Dioxide Balance of Wood
Substitution: Comparing Concrete- and Wood-Framed Buildings.” Mitigation and
Adaptation Strategies for Global Change 11 (3): 667–91. [Link]
Hanazato, Toshikazu, Kaori Fujita, Isao Sakamoto, Masahiro Inayama, and Yasuhiko Ohkura. 2004.
“Analysis of Earthquake Resistance of Five-Storied Timber Pagoda.” In . Vancouver, BC,
Canada.
Hankinson, R. L. 1921. “Investigation of Crushing Strength of Spruce at Varying Angles of Grain.”
Air Service Information Circular 3 (259): 130.
Henrichsen, Christoph, and Roland Bauer. 2004. Japan Culture of Wood: Buildings,
Objects,Techniques. 1 edition. Boston: Birkhauser.
Huixian, Liu, George W. Housner, Xie Lili, and He Duxin. 2002. “The Great Tangshan Earthquake
of 1976.” Report or Paper. California Institute of Technology.
[Link]
Inayama, Masahiro. 1991. “Wooden embedment theory and its application.” PhD, Tokyo, Japan:
University of Tokyo.
———. 1993. “Study on compression perpendicular to the grain in wood Part 4: analytic
functions for the relation between compression load and elastic deformation
perpendicular to the grain in wood.” In Proceedings of AIJ, 907–8. Tokyo.
International Energy Agency (IEA), and Global Alliance for Buildings and Construction
(GlobalABC). 2018. “2018 Global Status Report: Towards a Zero-Emission, Efficient and
Resilient Buildings and Construction Sector.” United Nationals Environment Programme.
[Link]
[Link].

109
 Demi Fang References
SMBT 2020

Jones, Craig, and Geoffrey Hammond. 2019. “Inventory of Carbon & Energy v3.0.”
[Link]
[Link].
Kellert, Stephen R., and Elizabeth F. Calabrese. 2015. “The Practice of Biophilic Design.” London:
Terrapin Bright LLC.
King, W. S., J. Y. Richard Yen, and Y. N. Alex Yen. 1996. “Joint Characteristics of Traditional
Chinese Wooden Frames.” Engineering Structures 18 (8): 635–44.
[Link]
Kitamori, A., T. Kataoka, and K. Komatsu. 2009. “Effect of Additional Length on Partial
Compression Perpendicular to the Grain of Wood.” Journal of Structural and
Construction Engineering 74: 642.
Komatsu, Kohei, Akihisa Kitamori, Kiho Jung, and Takuro Mori. 2009. “Estimation of the
Mechanical Properties of Mud Shear Walls Subjecting to Lateral Shear Force.” In
Proceedings of the 11th International Conference on Non-Conventional Materials and
Technologies. Bath, UK.
KT Innovations, thinkstep, and Autodesk. 2019. Tally(R) Life Cycle Assessment App.
[Link]
Lang, Elemer M., and Tamas Fodor. 2007. “Finite Element Analysis of Cross-Halved Joints for
Structural Composites.” Wood and Fiber Science 34 (2): 251–65.
Lenzen, M, and G Treloar. 2002. “Embodied Energy in Buildings: Wood versus Concrete—Reply
to Börjesson and Gustavsson.” Energy Policy 30 (3): 249–55.
[Link]
Li, Xiaowei, Junhai Zhao, Guowei Ma, and Wei Chen. 2015. “Experimental Study on the Seismic
Performance of a Double-Span Traditional Timber Frame.” Engineering Structures 98
(September): 141–50. [Link]
Miller, Laura Marjorie. 2017. “Mass Timber: Inside the Graceful, Sustainable, and Breathtaking
New Design Building.” The Magazine of the University of Massachusetts Amherst, May
24, 2017. [Link]
MTC Solutions. 2020. “Beam Hanger Design Guide.”
naturally:wood. 2016. “Design and Preconstruction of a Tall Wood Building, Brock Commons
Phase 1: Overview.”
Nordic Engineered Wood. 2013. “Nordic Lam Beams and Headers.”
Ogawa, Keita, Yasutoshi Sasaki, and Mariko Yamasaki. 2015. “Theoretical Modeling and
Experimental Study of Japanese ‘Watari-Ago’ Joints.” Journal of Wood Science 61 (5):
481–91. [Link]
Page, Mitchell. 2017. “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), 456–63. Cambridge, MA.
[Link]
Pomponi, Francesco, and Alice Moncaster. 2016. “Embodied Carbon Mitigation and Reduction
in the Built Environment – What Does the Evidence Say?” Journal of Environmental
Management 181 (October): 687–700. [Link]
Presinger, Clemens, and Bollinger+Grohmann. 2018. Karamba (version 1.2.2). Windows.
[Link]

110
 Demi Fang References
SMBT 2020

Que, Ze-li, Zhe-rui Li, Xiao-lan Zhang, Zi-ye Yuan, and Biao Pan. 2017. “Traditional Wooden
Buildings in China.” In Wood in Civil Engineering, edited by Giovanna Concu. InTech.
[Link]
Roche, Stéphane Nicolas. 2017. “Semi-Rigid Moment-Resisting Behavior of Multiple Tab-and-
Slot Joint for Freeform Timber Plate Structures.” PhD, Lausanne, Switzerland: Ecole
Polytechnique Federale de Lausanne. [Link]
Sathre, Roger, and Jennifer O’Connor. 2010. “Meta-Analysis of Greenhouse Gas Displacement
Factors of Wood Product Substitution.” Environmental Science & Policy 13 (2): 104–114.
Schmidt, Richard J., and Robert B. Mackay. 1997. “Timber Frame Tension Joinery.” MSc thesis,
Laramie, WY: University of Wyoming.
Schuchart. n.d. “The Bullitt Center.” Schuchart. Accessed May 3, 2020.
[Link]
Sebera, Václav, and Milan Šimek. 2010. “Finite Element Analysis of Dovetail Joint Made with the
Use of CNC Technology.” Acta Universitatis Agriculturae et Silviculturae Mendelianae
Brunensis 58 (5): 321–28. [Link]
Shanks, Jonathan, and Peter Walker. 2009. “Strength and Stiffness of All-Timber Pegged
Connections.” Journal of Materials in Civil Engineering 21 (1): 10–18.
[Link]
Shao, Benshun, Nicolette Lewis, Aliz Fischer, Yuli Huang, and Francois Lancelot. 2019.
“Parameters Identification for Wood Material (*MAT_143) and Its Application on the
Modeling of a Typical Timber Nuki Joint.” In . Shanghai, China.
Shiping, Hu. 1991. “The Earthquake‐Resistant Properties of Chinese Traditional Architecture.”
Earthquake Spectra 7 (3): 355–89. [Link]
Shope, Mitchell Grafton. 2016. “Strength Characterization of Wood to Wood Connections Using
Stress Field Analysis.” Thesis, Massachusetts Institute of Technology.
[Link]
Skidmore, Owings & Merrill, LLP. 2013. “Timber Tower Research Project: Initial Research Report.”
[Link]
Stern, Brenda G. 2018. “Minimizing Embodied Carbon in Multi-Material Structural Optimization
of Planar Trusses.” Thesis, Massachusetts Institute of Technology.
[Link]
“Structural Engineers 2050 Challenge.” n.d. Carbon Leadership Forum. Accessed December 15,
2019. [Link]
Suzuki, Yoshiyuki, and Masaki Maeno. 2006. “Structural Mechanism of Traditional Wooden
Frames by Dynamic and Static Tests.” Structural Control and Health Monitoring 13 (1):
508–22. [Link]
Takabayashi, Hiroki, Keita Kado, and Gakuhito Hirasawa. 2019. “Versatile Robotic Wood
Processing Based on Analysis of Parts Processing of Japanese Traditional Wooden
Buildings.” In Robotic Fabrication in Architecture, Art and Design 2018, edited by Jan
Willmann, Philippe Block, Marco Hutter, Kendra Byrne, and Tim Schork, 221–31. Springer
International Publishing.
Tanabashi, Ryo. 1960. “Earthquake Resistance of Traditional Japanese Wooden Structures” 40
(December): 1–15.
Tanahashi, Hideaki, and Yoshiyuki Suzuki. 2020. “Review on the Mechanical Models and
Formulations of Embedment of Traditional Timber Joints in Japan.” Japan Architectural
Review 3 (2): 148–64. [Link]

111
 Demi Fang References
SMBT 2020

Tannert, Thomas, Frank Lam, and Till Vallée. 2010. “Strength Prediction for Rounded Dovetail
Connections Considering Size Effects.” Journal of Engineering Mechanics 136 (3): 358–
66. [Link]
“The Bed.” n.d. Thuma. Accessed May 4, 2020. [Link]
Trussoni, Matthew, Evan Simatic, Christopher H. Raebel, and H. Peter Huttelmaier. 2015. “Life-
Cycle Assessment Comparison for Long-Span Cable and Truss Structural Systems: Case
Study.” Journal of Architectural Engineering 21 (1): 05014005.
[Link]
WoodWorks. n.d. “Promega Feynman Center, ‘The Crossroads.’” WoodWorks. Accessed May 3,
2020. [Link]
Wu, Ya Jie, Xiao Bin Song, and Lie Luo. 2017. “Experimental Investigation on the Seismic
Performance of a Chinese Traditional Wooden Pagoda.” Applied Mechanics and
Materials. [Link]
Wu, Yajie, Xiaobin Song, Xianglin Gu, and Lie Luo. 2018. “Dynamic Performance of a Multi-Story
Traditional Timber Pagoda.” Engineering Structures 159 (March): 277–85.
[Link]
Wu, Yajie, Xiaobin Song, and Kai Li. 2018. “Compressive and Racking Performance of
Eccentrically Aligned Dou-Gong Connections.” Engineering Structures 175 (November):
743–52. [Link]
Wu, Yajie, Xiaobin Song, and Carlos E. Ventura. 2019. “Modal Properties of a Model of a Chinese
Pagoda.” In Dynamics of Civil Structures, Volume 2, edited by Shamim Pakzad, 359–65.
Conference Proceedings of the Society for Experimental Mechanics Series. Springer
International Publishing.
Xie, Qifang, Long Wang, Lipeng Zhang, Weibing Hu, and Tiegang Zhou. 2018. “Seismic
Behaviour of a Traditional Timber Structure: Shaking Table Tests, Energy Dissipation
Mechanism and Damage Assessment Model.” Bulletin of Earthquake Engineering,
October. [Link]
Xue, Jianyang, Dan Xu, and Hailun Xia. 2018. “Experimental Study on Seismic Performance of
Through-Tenon Joints with Looseness in Ancient Timber Structures.” International
Journal of Architectural Heritage 0 (0): 1–13.
[Link]
Yeo, Sok Yee, Min-Fu Hsu, Kohei Komatsu, Yu-Lin Chung, and Wen-Shao Chang. 2016. “Shaking
Table Test of the Taiwanese Traditional Dieh-Dou Timber Frame.” International Journal
of Architectural Heritage 10 (5): 539–57. [Link]
Yeo, Sok-Yee, Kohei Komatsu, Min-Fu Hsu, and Zeli Que. 2016. “Mechanical Model for Complex
Brackets System of the Taiwanese Traditional Dieh-Dou Timber Structures.” Advances in
Structural Engineering 19 (1): 65–85. [Link]
Zwerger, Klaus. 2015. Wood and Wood Joints. 3rd ed. edition. Basel ; Boston: Birkhäuser.

112
Appendix 1: Analytical models for the Nuki joint
The code provided below was used to construct analytical models to match the Douglas fir
prototypes using the Maple symbolic computing environment. The code is the same for the
glulam prototypes but with different input parameters as detailed in Table 5. This code was
further developed by the author but primarily originated from collaborator Jan Brütting.
Part 1: Set up parameters and definitions
Note: some input parameters vary between the Douglas fir and glulam prototypes. Check Table
4 and Table 5 for respective parameters.
Input code Output
restart:

E__0 := 1.57e6; #psi 𝐸𝐸0 ∶= 1.57 × 106

E__90 := 0.9 * 40550; #psi 𝐸𝐸90 ∶= 36495.0
yield_strain := 0.018; 𝑦𝑦𝑦𝑦𝑦𝑦𝑦𝑦𝑦𝑦_𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 ∶= 0.018
Demi Fang Appendix 1: Analytical models for the Nuki
SMBT 2020 joint

mue := 0.2; 𝑚𝑚𝑚𝑚𝑚𝑚 ∶= 0.2

R := 0.07; # reduction in stiffness from elastic to plastic 𝑅𝑅 ∶= 0.07

Cw := 3.25; #inches 𝐶𝐶𝐶𝐶 ∶= 3.25

Cd := 3.25; #inches 𝐶𝐶𝐶𝐶 ∶= 3.25
Bd := 3.25; #inches 𝐵𝐵𝐵𝐵 ∶= 3.25

#l__e is the length of the short free end, which is 1" in all of our
prototypes ; l__c is the length of the long end and in literature is
taken as 1.5*Bd. The decay works across these lengths.
𝑙𝑙𝑒𝑒 ∶= 1
l__e := 1;
𝑙𝑙𝑐𝑐 ∶= 4.875
l__c := 1.5*Bd;

1.625
L := (Cd/2)/cos(theta); #length of beam that gets 𝐿𝐿 ≔
cos(𝜃𝜃)
compressed (distance to neutral axis)

1.625sin(𝜃𝜃)
delta := sin(theta)*L; #peak embedment length δ≔
cos(𝜃𝜃)

Z__theta := Bd*cos(theta); #new compressed beam depth 𝑍𝑍𝑡𝑡ℎ𝑒𝑒𝑒𝑒𝑒𝑒 ∶= 3.25 cos(𝜃𝜃)

0.5 sin(𝜃𝜃)
𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 ≔
strain := delta / Bd; cos(𝜃𝜃)

delta_yield := yield_strain * Bd; #delta at which the 𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑_𝑦𝑦𝑦𝑦𝑦𝑦𝑦𝑦𝑦𝑦 ∶= 0.05850

material yields (inches) 𝑡𝑡ℎ𝑒𝑒𝑒𝑒𝑒𝑒_𝑦𝑦𝑦𝑦𝑦𝑦𝑦𝑦𝑦𝑦 ∶
theta_yield := arctan(delta_yield / (Cd/2)); #theta at which the = 0.03598446008
material yields (radians)

#Hankinson's formula
E__theta := E__0*E__90/(E__0*(sin(3.1415927/2-theta))^2 +
E__90*(cos(3.1415927/2-theta))^2):

114
 Demi Fang Appendix 1: Analytical models for the Nuki
SMBT 2020 joint

Part 2: Calculate moment as a function of rotational displacement

This portion of code can be applied based on the input parameters defined in the previous
segment of code.
Input code Related
equation in
Section 3.2

#Embedments

#decay coefficients
decay__e := 6.5/Bd:
decay__c := 6.5/Bd:

#Ended embedment (short end)

f__e := delta*exp(-decay__e*x): (11)

#Continuous embedment (long end)

f__c := delta*exp(-decay__c*x): (11)

#ELASTIC Range

#compute areas
V1 := 0.5 * Cd/2 * delta: (12)
Ve1 := int(f__e, x=0..l__e): (13)
Vc1 := int(f__c, x=0..l__c): (14)

#compute corresponding moment arms

a1 := 2/3 * Cd/2: (19)
ae1 := Cd/2 + int(x*f__e, x=0..l__e)/Ve1: #use integration to find (20)
centroid location for moment arm
ac1 := Cd/2 + int(x*f__c, x=0..l__c)/Vc1:
(21)

#compute resultant forces

115
Demi Fang Appendix 1: Analytical models for the Nuki
SMBT 2020 joint

#N = volume * stiffness / shortened beam depth

Ne1 := (Bw * Ve1) * E__theta / Z__theta: (16)
Nc1 := (Bw * Vc1) * E__theta / Z__theta: (17)
N1 := (Bw * V1) * E__theta / Z__theta: (18)

M_el := 2*N1*a1 + Ne1*ae1 + Nc1*ac1: (51)

M_el := M_el + mue * N1 * Bd: # the extended length embedments (48)(49)(51)
(exponential curves) do not contribute to friction

#PLASTIC Range

delta_plastic := delta - delta_yield:

#Embedment from direct contact

Lp := Cd/2*(1 - delta_yield/delta): (22)

#compute areas
V2 := 0.5 * delta_yield * (Cd/2 - Lp): (24)
V3 := delta_yield * Lp: (25)
V4 := 0.5 * delta_plastic * Lp: (26)

#compute corresponding moment arms

a2 := (2/3)*(Cd/2 - Lp): (39)
a3 := Cd/2 - Lp/2: (40)
a4 := Cd/2 - Lp/3: (41)

#compute resultant forces

#N = volume * stiffness / shortened beam depth
N2 := (Bw * V2) * E__theta / Z__theta: (33)
N3 := (Bw * V3) * E__theta / Z__theta: (33)
N4 := (Bw * V4) * E__theta / Z__theta: (34)

116
Demi Fang Appendix 1: Analytical models for the Nuki
SMBT 2020 joint

M_dir_p := 2 * (N2*a2 + N3*a3 + R*N4*a4): Part of (52)

M_dir_p := M_dir_p + mue*(N2 + N3 + N4)*Bd: (48)(49)

#Embedment from indirect contact

Lpe := min(solve(f__e - delta_yield = 0, x), l__e): # (23)

horizontal distances at which embedment stress = yield stress
Lpc := min(solve(f__c - delta_yield = 0, x), l__c):
(23)

#End side

#compute areas
Ve2 := Lpe * delta_yield:
(27)
Ve3 := int(f__e, x=Lpe..l__e):
(28)
Ve4 := int(f__e - delta_yield, x=0..Lpe):
(29)

#compute corresponding moment arms

ae2 := Cd/2 + Lpe/2:
(42)
ae3 := Cd/2 + int(x*f__e, x=Lpe..l__e)/Ve3: #use integration to
(43)
find centroid location for moment arm
(44)
ae4 := Cd/2 + int(x*(f__e - delta_yield), x=0..Lpe)/Ve4:

#Continuous side

#compute areas
(30)
Vc2 := Lpc * delta_yield:
(31)
Vc3 := int(f__c, x=Lpc..l__c):
(32)
Vc4 := int(f__c - delta_yield, x=0..Lpc):

#compute corresponding moment arms

ac2 := Cd/2 + Lpc/2:
(45)
ac3 := Cd/2 + int(x*f__c, x=Lpc..l__c)/Vc3: #use integration to
find centroid location for moment arm (46)

117
 Demi Fang Appendix 1: Analytical models for the Nuki
SMBT 2020 joint

ac4 := Cd/2 + int(x*(f__c - delta_yield), x=0..Lpc)/Vc4: (47)

#compute resultant forces

Ne2 := (Bw* Ve2) * E__theta / Z__theta: (35)
Ne3 := (Bw* Ve3) * E__theta / Z__theta: (35)
Ne4 := (Bw* Ve4) * E__theta / Z__theta: (36)
Nc2 := (Bw* Vc2) * E__theta / Z__theta: (37)
Nc3 := (Bw* Vc3) * E__theta / Z__theta: (37)
Nc4 := (Bw* Vc4) * E__theta / Z__theta: (38)

M_embed_plastic := Ne2*ae2 + Ne3*ae3 + R*Ne4*ae4 + Nc2*ac2 + Nc3*ac3 Part of (52)

+ R*Nc4*ac4: #friction effects do not apply for indirect embedment

M_pl := M_dir_p + M_embed_plastic:

(52)
#compute rotational stiffnesses
pd_el := diff(M_el, theta):
pd_pl := diff(M_pl, theta):

Part 3: Assemble M as a function of theta for some range of theta

This example code inputs a beam width 𝐵𝐵𝑤𝑤 of 1”, which can be substituted with 1.5” for the
other beam width.
with(ArrayTools):

M1 := Array([0]):
rot1 := Array([0]):
diff1 := Array([0]):

#specify beam width

Bw := 1; #inches 𝐵𝐵𝐵𝐵 ∶= 1

#populate arrays of M and theta for a specified range of theta

for theta from 0.01 by 0.01 to 0.2 do

118
Demi Fang Appendix 1: Analytical models for the Nuki
SMBT 2020 joint

if delta <= delta_yield

then
Append(M1,M_el);
Append(rot1,theta);
Append(diff1,pd_el);
else
Append(M1,M_pl);
Append(rot1,theta);
Append(diff1,pd_pl);
end if;
end do:

#compute elastic stiffness

theta :=theta_yield/2:
pd_el; 76808.77400

#compute plastic stiffness

theta := theta_yield*3:
pd_pl; 25383.41518

119
Demi Fang Appendix 1: Analytical models for the Nuki
SMBT 2020 joint

120
 Demi Fang Appendix 2: Scans of tested prototypes
SMBT 2020

Appendix 2: Scans of tested prototypes

Douglas fir prototypes
This section includes scans of the beam elements of the Douglas fir prototypes after rotational
stiffness testing. The scans are oriented such that the ended end is at the top and the continuous
end at the bottom. Embedment is visible on the left and right edges.

121
Demi Fang Appendix 2: Scans of tested prototypes
SMBT 2020

Figure 52. Specimen 1 of the Douglas fir Nuki joint of Bw = 1", scanned after rotational stiffness testing.

122
Demi Fang Appendix 2: Scans of tested prototypes
SMBT 2020

Figure 53. Specimen 2 of the Douglas fir Nuki joint of Bw = 1", scanned after rotational stiffness testing.

123
Demi Fang Appendix 2: Scans of tested prototypes
SMBT 2020

Figure 54. Specimen 3 of the Douglas fir Nuki joint of Bw = 1", scanned after rotational stiffness testing.

124
 Demi Fang Appendix 2: Scans of tested prototypes
SMBT 2020

Figure 55. Specimen 1 of the Douglas fir Nuki joint of Bw = 1.5", scanned after rotational stiffness testing.

125
 Demi Fang Appendix 2: Scans of tested prototypes
SMBT 2020

Figure 56. Specimen 2 of the Douglas fir Nuki joint of Bw = 1.5", scanned after rotational stiffness testing.

126
 Demi Fang Appendix 2: Scans of tested prototypes
SMBT 2020

Figure 57. Specimen 3 of the Douglas fir Nuki joint of Bw = 1.5", scanned after rotational stiffness testing.

127
 Demi Fang Appendix 2: Scans of tested prototypes
SMBT 2020

Glulam prototypes
This section includes scans of the beam elements of some of the glulam prototypes after
rotational stiffness testing. Not all are shown because not all of the prototypes could be
disassembled by hand, but none of those depicted here were discarded as outliers in Section
4.3.2. The scans are oriented such that the ended end is at the top and the continuous end at
the bottom. Embedment is visible on the left and right edges.

128
Demi Fang Appendix 2: Scans of tested prototypes
SMBT 2020

Figure 58. Specimen 5 of the glulam Nuki joint of Bw = 1", scanned after rotational stiffness testing.

129
Demi Fang Appendix 2: Scans of tested prototypes
SMBT 2020

Figure 59. Specimen 6 of the glulam Nuki joint of Bw = 1", scanned after rotational stiffness testing.

130
Demi Fang Appendix 2: Scans of tested prototypes
SMBT 2020

Figure 60. Specimen 2 of the glulam Nuki joint of Bw = 1.5", scanned after rotational stiffness testing.

131
Demi Fang Appendix 2: Scans of tested prototypes
SMBT 2020

132
Appendix 3: Beam with partially rotational
restrained supports
This section will show the derivation of an indeterminate beam supported on both ends with a
support of a known partial rotational stiffness (intermediate between pin of 0 rotational stiffness
and a fixed support of infinite rotational stiffness; see Figure 61). The aim is to determine the
moment reactions at either end of the beam given the partial rotational stiffness at each support.
The moment distribution and deflected shape follow.
 Demi Fang Appendix 3: Beam with partially rotational
SMBT 2020 restrained supports

Figure 61. Comparison of uniformly loaded beam with varying boundary conditions.

Suppose a beam of length 𝐿𝐿 experiences a uniformly distributed load of 𝑞𝑞. At each end is a
support that provides upwards restraint and partial rotational stiffness of 𝑘𝑘 (in units of moment
per radians) in the elastic region.
𝑞𝑞𝑞𝑞
Due to symmetry, the vertical reaction forces are each known to equal . The moment reactions
2
at each support of magnitude 𝑀𝑀𝑅𝑅 are also known to be equivalent, i.e. −𝑀𝑀𝑅𝑅 = 𝑀𝑀(0) = 𝑀𝑀(𝐿𝐿), but
are still unknown. The free body diagram and internal force distributions are illustrated in Figure
62.

134
 Demi Fang Appendix 3: Beam with partially rotational
SMBT 2020 restrained supports

Figure 62. Free body diagram, shear V(x), moment M(x), and deflection w(x) diagrams for the beam
supported with partial rotational restraint.

Let 𝑥𝑥 represent the axis along the centroid of the beam. The internal shear within the beam 𝑉𝑉(𝑥𝑥)
can be expressed as follows:
𝑞𝑞𝑞𝑞
𝑉𝑉(𝑥𝑥) = − 𝑞𝑞𝑞𝑞 (65)
2
It follows that the internal bending moment within the beam 𝑀𝑀(𝑥𝑥) is:
𝑞𝑞𝑞𝑞 𝑞𝑞
𝑀𝑀(𝑥𝑥) = � 𝑉𝑉(𝑥𝑥)𝑑𝑑𝑑𝑑 = −𝑀𝑀𝑅𝑅 + 𝑥𝑥 − 𝑥𝑥 2 (66)
2 2
To use differential equations to solve for the beams, boundary conditions must first be stated.
Let 𝑤𝑤(𝑥𝑥) represent the deflection of the beam perpendicular to its axis (positive upwards), and
𝜑𝜑(𝑥𝑥) = 𝑤𝑤′(𝑥𝑥) represent the angle of deflection. The following boundary conditions arise out of
the supports and out of symmetry:
𝑤𝑤(0) = 𝑤𝑤(𝐿𝐿) = 0 (67)
𝑀𝑀𝑅𝑅
𝑤𝑤 ′ (0) = 𝑤𝑤′(𝐿𝐿) = − (68)
𝑘𝑘
𝐿𝐿
𝑤𝑤 ′ � � = 0 (69)
2
Assuming Euler-Bernoulli beam theory applies, we can set the expression for 𝑀𝑀(𝑥𝑥) equal to
𝐸𝐸𝐸𝐸𝐸𝐸′′(𝑥𝑥):

135
 Demi Fang Appendix 3: Beam with partially rotational
SMBT 2020 restrained supports

𝑞𝑞𝑞𝑞 𝑞𝑞
𝐸𝐸𝐸𝐸𝐸𝐸′′(𝑥𝑥) = −𝑀𝑀𝑅𝑅 + 𝑥𝑥 − 𝑥𝑥 2 (70)
2 2
𝑞𝑞𝑞𝑞 𝑞𝑞
𝐸𝐸𝐸𝐸𝐸𝐸′(𝑥𝑥) = 𝐶𝐶 − 𝑀𝑀𝑅𝑅 𝑥𝑥 + 𝑥𝑥 2 − 𝑥𝑥 3 (71)
4 6
The two boundary conditions (68) and (69) can be substituted into (71) to solve for both the
integration constant 𝐶𝐶 and the reaction moment 𝑀𝑀𝑅𝑅 .
𝑀𝑀𝑅𝑅
𝐸𝐸𝐸𝐸𝑤𝑤 ′ (0) = −𝐸𝐸𝐸𝐸 = 𝐶𝐶 (72)
𝑘𝑘
𝐿𝐿 𝑀𝑀𝑅𝑅 𝐿𝐿 𝑞𝑞𝑞𝑞 𝐿𝐿 2 𝑞𝑞 𝐿𝐿 3
𝐸𝐸𝐸𝐸𝑤𝑤 ′ � � = 0 = 𝐶𝐶 − + � � − � � (73)
2 2 4 2 6 2
𝑀𝑀𝑅𝑅 𝑀𝑀𝑅𝑅 𝐿𝐿 𝑞𝑞𝐿𝐿3 𝑞𝑞𝐿𝐿3 𝑀𝑀𝑅𝑅 𝐿𝐿 𝑞𝑞𝐿𝐿3
𝐶𝐶 = −𝐸𝐸𝐸𝐸 = − + = − (74)
𝑘𝑘 2 16 48 2 24
𝐸𝐸𝐸𝐸 𝐿𝐿 𝑞𝑞𝐿𝐿3
𝑀𝑀𝑅𝑅 �− − � = −
𝑘𝑘 2 24
𝑞𝑞𝐿𝐿3
𝑀𝑀𝑅𝑅 =
𝐸𝐸𝐸𝐸 𝐿𝐿
24 � + � (75)
𝑘𝑘 2

The expression for 𝑀𝑀𝑅𝑅 in equation (75) is used in Section 5.3.3 to determine the moment
distribution of a beam supported on both ends with a joint of known rotational stiffness 𝑘𝑘.
Another value of interest is the maximum deflection in the beam. Due to the symmetric nature
of the loading and support conditions, the maximum deflection is known to occur at midspan.
𝐿𝐿
The value of 𝑤𝑤 � � is thus the maximum deflection of interest. We can integrate equation (71),
2
apply boundary conditions from equation (67), and substitute integration constant C from
equation (74) to find 𝑤𝑤(𝑥𝑥):
𝑀𝑀𝑅𝑅 2 𝑞𝑞𝑞𝑞 3 𝑞𝑞 4
𝐸𝐸𝐸𝐸𝐸𝐸(𝑥𝑥) = 𝐶𝐶0 + 𝐶𝐶𝐶𝐶 − 𝑥𝑥 + 𝑥𝑥 − 𝑥𝑥
2 12 24
𝐸𝐸𝐸𝐸𝐸𝐸(0) = 0 = 𝐶𝐶0
𝑀𝑀𝑅𝑅 2 𝑞𝑞𝑞𝑞 3 𝑞𝑞
𝐸𝐸𝐸𝐸𝐸𝐸(𝑥𝑥) = 𝐶𝐶𝐶𝐶 − 𝑥𝑥 + 𝑥𝑥 − 𝑥𝑥 4
2 12 24
𝑀𝑀𝑅𝑅 𝑀𝑀𝑅𝑅 2 𝑞𝑞𝑞𝑞 3 𝑞𝑞 4
𝐸𝐸𝐸𝐸𝐸𝐸(𝑥𝑥) = −𝐸𝐸𝐸𝐸 𝑥𝑥 − 𝑥𝑥 + 𝑥𝑥 − 𝑥𝑥
𝑘𝑘 2 12 24
1 𝑀𝑀𝑅𝑅 𝑀𝑀𝑅𝑅 2 𝑞𝑞𝑞𝑞 3 𝑞𝑞
𝑤𝑤(𝑥𝑥) = �−𝐸𝐸𝐸𝐸 𝑥𝑥 − 𝑥𝑥 + 𝑥𝑥 − 𝑥𝑥 4 � (76)
𝐸𝐸𝐸𝐸 𝑘𝑘 2 12 24

where 𝑀𝑀𝑅𝑅 is given by equation (75).

Finally, we can find midspan deflection:

136
 Demi Fang Appendix 3: Beam with partially rotational
SMBT 2020 restrained supports

𝐿𝐿 1 𝑀𝑀𝑅𝑅 𝐿𝐿 𝑀𝑀𝑅𝑅 𝐿𝐿2 𝑞𝑞𝑞𝑞 𝐿𝐿3 𝑞𝑞 𝐿𝐿4

𝑤𝑤 � � = �−𝐸𝐸𝐸𝐸 − + − �
2 𝐸𝐸𝐸𝐸 𝑘𝑘 2 2 4 12 8 24 16
𝐿𝐿 𝐿𝐿 𝑀𝑀𝑅𝑅 𝑀𝑀𝑅𝑅 𝐿𝐿 𝑞𝑞𝐿𝐿3
𝑤𝑤 � � = �−𝐸𝐸𝐸𝐸 − + � (77)
2 𝐸𝐸𝐸𝐸 2𝑘𝑘 8 128
where 𝑀𝑀𝑅𝑅 is given by equation (75).

137
Demi Fang Appendix 3: Beam with partially rotational
SMBT 2020 restrained supports

138
Appendix 4: Beam hanger volume estimates
In this section, details for selecting and calculating quantities for beam hanger products used in
Section 5 are provided. The final volume quantities calculated in this appendix are used in Table
28, Table 29, Table 30, and Table 31.
Ricon S VS 140x60
The 10 ft span required a shear capacity of at least 2.2 k. Assuming SPF glulam, to provide a
shear capacity of 2.44 k, a single Ricon S VS 140x60 beam hanger was used with 7 fasteners on
the primary member and 7 fasteners on the secondary member. The volume calculation of this
beam hanger design is presented in Table 33 and is input into the total quantities of the 10 ft
span supported by beam hangers (Table 28).
 Demi Fang Appendix 4: Beam hanger volume estimates
SMBT 2020

Table 33. Quantities calculation for a Ricon S VS 140x60 beam hanger.

Connector Fastener: VG CSK Fastener: VG CSK

plate* 5/16” x 3-1/8” 5/16” x 6-1/4”
2-
𝑤𝑤 in. 𝑑𝑑𝑡𝑡ℎ𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 5/16 in. 𝑑𝑑𝑡𝑡ℎ𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 5/16 in.
3/8
5-
ℎ in. 𝑑𝑑ℎ𝑒𝑒𝑒𝑒𝑒𝑒 5/8 in. 𝑑𝑑ℎ𝑒𝑒𝑒𝑒𝑒𝑒 5/8 in.
1/2
𝑡𝑡 3/16 in. 𝐿𝐿𝑡𝑡ℎ𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 2-1/2 in. 𝐿𝐿𝑡𝑡ℎ𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 5-5/8 in.
𝐿𝐿 3-1/8 in. 𝐿𝐿 6-1/4 in.
Volume steel
2.45 in.3 0 in.3 0 in.3
plate
Volume steel
0 in.3 0.256 in.3 0.495 in.3
wire rod**
Qty per
2 7 7
connector
Qty connectors
2 2 2
per beam
Total volume
steel plate per 9.8 in.3 0 in.3 0 in.3 9.8 in.3
beam
Total volume
steel wire rod 0 in.3 3.58 in.3 6.93 in.3 10.5 in.3
per beam
*Cutouts, fillets, and knob are approximated to cancel each other out and are omitted from
volume calculation.
𝑑𝑑𝑡𝑡ℎ𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 2 1 𝑑𝑑ℎ𝑒𝑒𝑒𝑒𝑒𝑒 2
**Volume of fasteners estimated as 𝜋𝜋 � � 𝐿𝐿𝑡𝑡ℎ𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 + 𝜋𝜋 � � (𝐿𝐿 − 𝐿𝐿𝑡𝑡ℎ𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 ).
2 3 2

Ricon S VS 200x80
The 15 ft span required a shear capacity of at least 4.95 k. Assuming SPF glulam, to provide a
shear capacity of 6.89 k, a single Ricon S VS 200x80 beam hanger was used with 16 fasteners on
the primary member and 16 fasteners on the secondary member. The volume calculation of this
beam hanger design is presented in Table 34 and is input into the total quantities of the 15 ft
span supported by beam hangers (Table 29).

140
 Demi Fang Appendix 4: Beam hanger volume estimates
SMBT 2020

Table 34. Quantities calculation for a Ricon S VS 200x80 beam hanger.

Connector Fastener: VG CSK Fastener: VG CSK

plate* 3/8” x 4” 3/8” x 7-7/8”
3-
𝑤𝑤 in. 𝑑𝑑𝑡𝑡ℎ𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 3/8 in. 𝑑𝑑𝑡𝑡ℎ𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 3/8 in.
1/8
7-
ℎ in. 𝑑𝑑ℎ𝑒𝑒𝑒𝑒𝑒𝑒 3/4 in. 𝑑𝑑ℎ𝑒𝑒𝑒𝑒𝑒𝑒 3/4 in.
7/8
2- 5-
𝑡𝑡 3/16 in. 𝐿𝐿𝑡𝑡ℎ𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 in. 𝐿𝐿𝑡𝑡ℎ𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 in.
1/2 5/8
7-
𝐿𝐿 4 in. 𝐿𝐿 in.
7/8
Volume steel
4.61 in.3 0 in.3 0 in.3
plate
Volume steel wire
0 in.3 0.497 in.3 0.953 in.3
rod**
Qty per
2 16 16
connector
Qty connectors
2 2 2
per beam
Total volume
steel plate per 18.5 in.3 0 in.3 0 in.3 18.5 in.3
beam
Total volume
steel wire rod per 0 in.3 15.9 in.3 30.5 in.3 46.4 in.3
beam
*Cutouts, fillets, and knob are approximated to cancel each other out and are omitted from
volume calculation.
𝑑𝑑𝑡𝑡ℎ𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 2 1 𝑑𝑑ℎ𝑒𝑒𝑒𝑒𝑒𝑒 2
**Volume of fasteners estimated as 𝜋𝜋 � 2
� 𝐿𝐿𝑡𝑡ℎ𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 + 3 𝜋𝜋 � 2
� (𝐿𝐿 − 𝐿𝐿𝑡𝑡ℎ𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 )

MEGANT 310x100
The 20 ft span required a shear capacity of at least 8.80 k. Assuming SPF glulam, to provide a
shear capacity of 9.28 k, a MEGANT 310x100 beam hanger was used with 34 fasteners.
Compared to the Ricon S VS line, the MEGANT line has several more components. Aluminum
components and steel wire rod components are separated out into Table 35 and Table 36,
respectively. Hex nuts and washers were omitted from volume calculations. The final quantities
are input into the total quantities of the 20 ft span supported by beam hangers (Table 30).
141
Demi Fang Appendix 4: Beam hanger volume estimates
SMBT 2020

Table 35. Aluminum quantities calculation for a MEGANT 310x100 beam hanger.

Connector plate* Clamping jaws**

𝑤𝑤 4 in. 𝑤𝑤𝑗𝑗 4 in.
𝑤𝑤2 7/8 in. ℎ𝑗𝑗 2 in.
ℎ3 9-7/8 in. 𝑡𝑡𝑗𝑗 1-9/16 in.
ℎ𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜 3/4 in. ℎ𝑗𝑗,𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔 3/4 in.
𝑡𝑡 1/2 in. 𝑡𝑡𝑗𝑗,𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔 1 in.
Volume aluminum per
30.74 in.3 9.50 in.3
component
Qty per connector 2 2
Qty connectors per beam 2 2
Total volume aluminum per
123.0 in.3 38.0 in.3 161.0 in.3
beam
*Volume approximated as ℎ3 𝑤𝑤𝑤𝑤 + (ℎ3 − 2ℎ𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜 )(3𝑤𝑤2 )𝑡𝑡 . Holes for fasteners omitted from
calculation.
**Volume approximated as ℎ𝑗𝑗 𝑤𝑤𝑗𝑗 𝑡𝑡𝑗𝑗 − ℎ𝑗𝑗,𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔 𝑤𝑤𝑗𝑗 𝑡𝑡𝑗𝑗,𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔 .

Table 36. Steel wire rod quantities calculation for a MEGANT 310x100 beam hanger.

Threaded Fastener: VG CSK 5/16” x 6-

rods 1/4” *
𝑑𝑑 3/4 in. 𝑑𝑑𝑡𝑡ℎ𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 5/16 in.
9-
𝐿𝐿 in. 𝑑𝑑ℎ𝑒𝑒𝑒𝑒𝑒𝑒 5/8 in.
7/8
𝐿𝐿𝑡𝑡ℎ𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 5-5/8 in.
𝐿𝐿 6-1/4 in.
Volume steel wire rod per
5.91 in.3 0.495 in.3
component
Qty per connector 2 34
Qty connectors per beam 2 2
Total volume steel wire rod per
23.6 in.3 33.7 in.3 57.3 in.3
beam
𝑑𝑑𝑡𝑡ℎ𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 2 1 𝑑𝑑ℎ𝑒𝑒𝑒𝑒𝑒𝑒 2
*Volume of fasteners estimated as 𝜋𝜋 � 2
� 𝐿𝐿𝑡𝑡ℎ𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 + 3 𝜋𝜋 � 2
� (𝐿𝐿 − 𝐿𝐿𝑡𝑡ℎ𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 ).

142
 Demi Fang Appendix 4: Beam hanger volume estimates
SMBT 2020

Ricon XL 390x80
The 25 ft span required a shear capacity of at least 13.75 k. Assuming SPF glulam, to provide a
shear capacity of 15.5 k, a single Ricon XL 390x80 beam hanger was used with 30 fasteners on
the primary member and 30 fasteners on the secondary member. The volume calculation of this
beam hanger design is presented in Table 37 and is input into the total quantities of the 25 ft
span supported by beam hangers (Table 31).
Table 37. Quantities calculation for a Ricon XL 390x80 beam hanger.

Connector plate + Fastener: VG CSK Fastener: VG CSK

block* 3/8” x 4” 3/8” x 7-7/8”
3-
𝑤𝑤 in. 𝑑𝑑𝑡𝑡ℎ𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 3/8 in. 𝑑𝑑𝑡𝑡ℎ𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 3/8 in.
1/8
13-
ℎ in. 𝑑𝑑ℎ𝑒𝑒𝑒𝑒𝑒𝑒 3/4 in. 𝑑𝑑ℎ𝑒𝑒𝑒𝑒𝑒𝑒 3/4 in.
3/8
2- 5-
𝑡𝑡 3/16 in. 𝐿𝐿𝑡𝑡ℎ𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 in. 𝐿𝐿𝑡𝑡ℎ𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 in.
1/2 5/8
3- 7-
𝑤𝑤𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 in. 𝐿𝐿 4 in. 𝐿𝐿 in.
1/8 7/8
ℎ𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 2 in.
𝑡𝑡𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 1 in.
Volume steel
14.09 in.3 0 in.3 0 in.3
plate
Volume steel
0 in.3 0.497 in.3 0.953 in.3
wire rod**
Qty per
2 30 30
connector
Qty
connectors 2 2 2
per beam
Total volume
steel plate per 56.3 in.3 0 in.3 0 in.3 56.3 in.3
beam
Total volume
steel wire rod 0 in.3 29.8 in.3 57.2 in.3 87.0 in.3
per beam
*Cutouts, fillets, and knob are approximated to cancel each other out and are omitted from
volume calculation. Volume of unit estimated as 𝑤𝑤ℎ𝑡𝑡 + 𝑤𝑤𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 ℎ𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 𝑡𝑡𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 .
𝑑𝑑𝑡𝑡ℎ𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 2 1 𝑑𝑑ℎ𝑒𝑒𝑎𝑎𝑎𝑎 2
**Volume of fasteners estimated as 𝜋𝜋 � 2
� 𝐿𝐿𝑡𝑡ℎ𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 + 3 𝜋𝜋 � 2
� (𝐿𝐿 − 𝐿𝐿𝑡𝑡ℎ𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 )

143
Demi Fang Appendix 4: Beam hanger volume estimates
SMBT 2020

144