This action might not be possible to undo. Are you sure you want to continue?
School of Engineering
Final Year Project
Structural Analysis of Geodesic
Domes
Author:
Marek Kubik
Supervisor:
Charles Augarde
April 29, 2009
Abstract
A project to house 40 families in the Maharashtra region of India which began
in January 2005 was halted shortly after it commenced due to concerns over
the loading applied to a series of geodesic domes which form a large portion of
the whole complex. Vigyan Ashram, the organisation that manufactures the
domes, issued a request for the development of an aﬀordable computer program
that would allow them to model the structural response of the domes. The
ensuing research into the design of the geodesic domes and the development of
a spreadsheet based ﬁnite element package are the subject of this dissertation.
Acknowledgements
I wish to thank all at Vigyan Ashram, particularly Yogesh Kulkarni and Ashok
Mathur, for their assistance with gathering information relevant to this project.
I would also like to express my gratitude to Colin Wintrip and Steven Richard
son, whose assistance with fabricating and testing the joints was greatly ap
preciated. Finally, I wish to thank Dr. Charles Augarde, my supervisor, who
has provided me with invaluable guidance and advice throughout this research
project.
i
Project plan
Introduction
A project to house 40 families in the Maharashtra region of India which began in January
2005 was halted shortly after it commenced due to concerns over the loading applied to
a series of geodesic domes which form a large portion of the whole complex. The domes
had soil packed over them, a condition which was never anticipated in the original design.
The need for a method to model the structural response of the geodesic or “Pabal” dome
was highlighted by Vigyan Ashram, the Non Governmental Organisation (NGO) that
manufactures the domes; this dissertation describes the research into the Pabal dome and
the development of a bespoke ﬁnite element analysis (FEA) package capable of running
in Microsoft Excel.
Objectives
Vigyan Ashram locally produces the geodesic domes as doityourself kits for the lower
middle class of both rural and urban populations. The original design was adopted in
the aftermath of the 1993 Killari earthquake, aiming to provide durable, low cost housing
capable of withstanding the earthquakes, rains and winds of India for those that lost their
homes.
Of the 120 geodesic dome kits supplied to the Water Bank housing project, 40 were
planned to be subterranean and were therefore aﬀected. Vigyan Ashram, together with
another NGO, engINdia, asked for research into the geodesic dome’s current incarnation
to be carried out. The Water Bank project highlighted the need for a method of modelling
the structural response of geodesic domes, as an assessment of the loading encountered
by the domes would allow recommendations to be made as to how the design could be
adapted to accommodate the expected loading.
Vigyan Ashram’s main desire was for a structural analysis package without the asso
ciated licensing costs, to allow their science and technology centre to assess the geodesic
dome’s structural response inhouse. This would provide signiﬁcant long term beneﬁts
for Vigyan Ashram, removing the need for outsider aid with such design problems in the
future. If appropriate, students at Vigyan Ashram would also use the developed method
to further their understanding of structural behaviour.
It was hoped that a reliable method of analysis would increase the number of potential
uses of the structure and hence its market demand, beneﬁting both the community and
the business that manufactures the dome, which was founded and run by an exstudent
of Vigyan Ashram.
The objectives of this project may be summarised as follows.
1. The development of a method of ﬁnite element analysis for geodesic dome structures
ii
using Microsoft Excel 2003. The program needed to provide a user friendly format
that gave the user the opportunity to deﬁne a geometry, select appropriate material
properties and apply a variety of loading conditions. It was required to output
information about nodal displacements and forces in the elements.
2. The assessment of the dome joints used in the structure to determine their design
limits. Based on these ﬁndings, a relationship between material parameters and the
ultimate failure strength of the connections was to be developed.
3. The development of an additional spreadsheet that would allow the user to calculate
all the relevant dome fabrication details  how many diﬀerent struts are required,
how long they are, how many bolts are required for suﬃcient connection strength
etc.
Project timeline
This project was to be carried out over a period of 28.5 weeks  from the end of September
through to midMay. A Gantt chart breakdown of the project plan may be found in
Appendix A, with all the major project deadlines highlighted.
Work was broadly categorised into three phases.
• Phase I: Initial research and planning; lasting from September through to January,
culminating in the submission of the literature review. In this phase, background
reading and discussion with Vigyan Ashram would lead to the development of a list
of objectives and a project plan. Additionally, preparation work for the later tension
testing work would be scheduled, including the submission of a risk assessment and
the production of engineering drawings to fabricate the required dome joints.
• Phase II: Data collection; overlapping with Phase I, from December to the end of
April. Here, experimental work would be carried out and the spreadsheets devel
oped, culminating in the submission of a draft version of the ﬁnal report.
• Phase III: Project conclusions; from April until the concluding oral exam in May.
Upon return of the draft report from the project supervisor, modiﬁcations would be
made and the ﬁnal project submitted on the 29th of April. Additionally, a poster
design for display in the department would be produced and preparation for the
oral exams would occur in this period.
iii
Contents
Abstract i
Acknowledgements i
Project Plan iii
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii
Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii
Project timeline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii
Table of Contents iv
List of Tables vi
List of Figures vii
1 Introduction 1
2 Literature review 1
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
2.2 Pabal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
2.2.1 Earthquake and aftermath . . . . . . . . . . . . . . . . . . . . . . . 2
2.2.2 Vigyan Ashram and engINdia . . . . . . . . . . . . . . . . . . . . . 3
2.2.3 The Water Bank project . . . . . . . . . . . . . . . . . . . . . . . . 4
2.2.4 Structural analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.3 Geodesic domes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.3.1 Buckminster Fuller . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.3.2 Deﬁnition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.3.3 Explanation of behaviour . . . . . . . . . . . . . . . . . . . . . . . . 8
2.3.4 Strengths and weaknesses . . . . . . . . . . . . . . . . . . . . . . . 8
2.3.5 The Pabal dome . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.4 Braced dome analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.4.1 Idealisations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.4.2 Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.4.3 Previous work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
3 Theory 12
3.1 Stiﬀness method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
3.2 Solution methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
3.3 Post processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3.4 Lateral earth pressure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
iv
4 Geodesic dome analysis 18
4.1 Introduction to Strand7 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
4.2 Coordinate generation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
4.3 Loading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
4.3.1 Seismic and impact loads . . . . . . . . . . . . . . . . . . . . . . . . 20
4.3.2 Snow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
4.3.3 Soil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
4.3.4 Wind . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
4.3.5 Self weight and imposed loads . . . . . . . . . . . . . . . . . . . . . 23
4.4 Assumptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
4.5 Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
4.5.1 Load Case I: Self weight . . . . . . . . . . . . . . . . . . . . . . . . 26
4.5.2 Load Cases II & III: Soil . . . . . . . . . . . . . . . . . . . . . . . . 26
4.5.3 Load Case IV: Wind . . . . . . . . . . . . . . . . . . . . . . . . . . 27
4.5.4 Load Case V: Imposed . . . . . . . . . . . . . . . . . . . . . . . . . 28
4.5.5 Combination load cases . . . . . . . . . . . . . . . . . . . . . . . . . 28
4.5.6 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
4.5.7 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
5 Joint testing 29
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
5.2 Pabal disc joints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
5.3 Dimensional analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
5.4 Tension testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
5.4.1 Joint fabrication . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
5.4.2 Apparatus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
5.4.3 Risk assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
5.4.4 Method statement . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
5.4.5 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
5.4.6 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
5.5 Stress analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
5.5.1 Model development . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
5.5.2 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
5.5.3 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
6 Spreadsheet development 40
6.1 Package structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
6.2 FEA_Input . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
6.3 FEA_Solver . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
6.3.1 Stiﬀness matrix assembly . . . . . . . . . . . . . . . . . . . . . . . . 41
v
6.3.2 Solver . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
6.4 FEA_Output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
6.4.1 Postprocessing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
6.4.2 Calculations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
6.5 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
7 Conclusions 48
7.1 Further work recommendations . . . . . . . . . . . . . . . . . . . . . . . . 48
References 50
A Project Gantt chart 53
B Joint forcedisplacement curves 53
C Wind loading calculations 55
D Joint engineering drawing 56
E Sample Excel calculation 57
F Risk Assessment and COSSH form 61
List of Tables
1 Element and joint properties . . . . . . . . . . . . . . . . . . . . . . . . . . 27
2 Black Cotton soil properties . . . . . . . . . . . . . . . . . . . . . . . . . . 28
3 Strand analysis results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
4 Dimensional analysis: step 1 . . . . . . . . . . . . . . . . . . . . . . . . . . 32
5 Dimensional analysis: step 2 . . . . . . . . . . . . . . . . . . . . . . . . . . 32
6 Dimensional analysis: step 3 . . . . . . . . . . . . . . . . . . . . . . . . . . 32
7 Convergence test results . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
8 Nonlinear joint analysis results . . . . . . . . . . . . . . . . . . . . . . . . 39
9 Comparison of theoretical and practical joint capacities . . . . . . . . . . . 39
10 Spreadsheet based FEA results . . . . . . . . . . . . . . . . . . . . . . . . 47
vi
List of Figures
1 Isosahedron . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2 Great circles [Morgan, 1981] . . . . . . . . . . . . . . . . . . . . . . . . . . 6
3 Class I subdivision [Motro, 1984] . . . . . . . . . . . . . . . . . . . . . . . 7
4 Class II subdivision [Motro, 1984] . . . . . . . . . . . . . . . . . . . . . . . 7
5 Example subdivision frequencies [Motro, 1984] . . . . . . . . . . . . . . . 7
6 Pabal dome . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
7 Brown University’s 2D FE program . . . . . . . . . . . . . . . . . . . . . . 11
8 DOF for a pin jointed element in a local coordinate system . . . . . . . . . 14
9 Assembly of the structure stiﬀness matrix from element stiﬀness matrices. . 15
10 Structure stiﬀness matrix bandwidth . . . . . . . . . . . . . . . . . . . . . 16
11 Global coordinate system used for analysis . . . . . . . . . . . . . . . . . . 19
12 Horizontal and vertical stresses acting on a subterranean dome (after Landry
[2002]) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
13 Regional soil deposits of India (after Shroﬀ and Shah [2003]) . . . . . . . . 22
14 External wind pressure coeﬃcients for a spherical structure (from Table
18, Bhandari et al. [1987]) . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
15 Example area division for a node . . . . . . . . . . . . . . . . . . . . . . . 24
16 Soil loads considered . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
17 Close up of Pabal dome joint . . . . . . . . . . . . . . . . . . . . . . . . . . 30
18 Uniaxial loading of bolt array . . . . . . . . . . . . . . . . . . . . . . . . . 31
19 Denison loading machine . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
20 Test rigs for joints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
21 Tension test results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
22 Mesh convergence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
23 Sketches of material stressstrain relationships . . . . . . . . . . . . . . . . 37
24 Plot of relationship between dimensionless variables . . . . . . . . . . . . . 38
25 Flow chart of design process . . . . . . . . . . . . . . . . . . . . . . . . . . 40
26 Calculation of project member lengths . . . . . . . . . . . . . . . . . . . . 42
27 Assembly of the global stiﬀness matrix . . . . . . . . . . . . . . . . . . . . 42
28 Intermediate matrix assembly step . . . . . . . . . . . . . . . . . . . . . . 43
29 Structure stiﬀness matrix assembly using SUMPRODUCT . . . . . . . . . 44
30 Calculation of member axial forces . . . . . . . . . . . . . . . . . . . . . . 45
31 Project Gantt chart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
32 Stress analysis forcedisplacement plot for 1.5mm disc . . . . . . . . . . . . 53
33 Stress analysis forcedisplacement plot for 2.0mm disc . . . . . . . . . . . . 54
34 Stress analysis forcedisplacement plot for 3.0mm disc . . . . . . . . . . . . 54
vii
1 Introduction
The project plan included in the prelude of this report deﬁnes the project objectives
agreed with Vigyan Ashram and the expected timeline in which they were to be achieved.
The literature review provides context and background information to justify the need
for this work and the theory section summarises the necessary information required to
understand the FEA method developed in Excel.
The main body of the report covers three main aspects of the project.
1. The development of a geodesic dome model and the subsequent analysis of the
dome using Strand7, a commercial FEA package, which provided an early basis
for gathering results and a benchmark against which to compare the Excel based
method.
2. The development of a relationship to predict the failure strength of the joints from
the material properties and geometry, using a combination of physical tension testing
and stress analysis.
3. An overview of the ﬁnite element (FE) method developed in Excel.
In the ﬁnal part of the report, the success of meeting the objectives outlined in the
project plan is discussed and conclusions are drawn, with recommendations for further
work in this area.
2 Literature review
2.1 Introduction
The purpose of this literature review is to give context to the relevance and necessity of
the work undertaken in this dissertation, and to introduce the concepts and deﬁnitions
that will aid the understanding of the rest of the work herein.
Firstly, a brief history of the geodesic dome’s role in the reconstruction and rehabil
itation in wake of the 1993 Killari earthquake is given. The NGO that has ultimately
continued to locally produce the dome kits to the present day, Vigyan Ashram, is then
introduced. The work halted at the Water Bank project in 2005 is used to demonstrate
the need for an assessment of the loading on the geodesic domes.
The second part of the literature review examines the deﬁnition of a geodesic dome
and looks brieﬂy at its history. The strengths and weaknesses of geodesic domes are
discussed, and the modiﬁed version of the dome used by Vigyan Ashram is introduced.
Finally, the possible methods of braced dome analysis are reviewed and the decision
to produce an Excel based program is justiﬁed. Previous work in this area is reviewed,
1
and the shortcomings are used to justify Vigyan Ashram’s need for a bespoke structural
analysis package.
2.2 Pabal
2.2.1 Earthquake and aftermath
The following description of the earthquake and aftermath are based upon extracts from
the account of Professor Horst Rolly [Rolly, 2007], who lived in the region at the time of
the quake.
In the early hours of the morning of the 30th September 1993, the Killari earthquake
struck the Maharashtra province of India. The ﬁrst quake lasted for 40 to 50 seconds,
survivors recounting that houses “swayed like a cradle” before they caved in and buried
people underneath. Animals, traditionally staying in rooms attached to houses, met the
same fate. The magnitude of this disaster became apparent only days later; 7,928 human
lives lost, a further 16,000 were injured and more than 15,800 livestock were killed. The
damage was total in 52 villages of the Latur and Osmanabad districts, but the eﬀects of
the disaster were felt in more than 2500 villages in the 11 neighbouring districts.
Real poverty proved to be a blessing as the landless poor lived in thatchedroof build
ings made of straw and reeds or split bamboo, ﬁnished with mudplaster. These houses
performed extremely well, suﬀering only minor cracks to the mud plaster walls. The
structures that suﬀered most were those constructed of thick masonry to provide better
insulation against the summer heat. Local builders reported that “. . . the failure of stone
masonry during the earthquake was largely because of the excessive wall thickness de
manded by the Maratha households for thermal comfort and storage of valuables within
the thickness of the walls.” No oﬃcial engineering standards existed in rural Maharashtra
before the quake; often boulders were piled upon shallow foundations, bonded only with
minimal cementing. In normal conditions these constructions are relatively stable, but on
the soft Black Cotton soil of the region they collapsed easily when the earth shook.
No early warning system for such a natural disaster was in place and no comprehensive
disaster management plan existed before the quake. Regardless, the civic response was
spontaneous. Ablebodied survivors assisted the injured and dug in debris to rescue the
living and recover the dead with their bare hands as initially no excavation equipment
was available. The Maharashtra state government in Bombay responded to the news im
mediately by sending helpers from civic bodies and doctors from neighbouring districts
along with supplies. The Indian armed forces were brought in to aid with the rescue and
relief operation, with some 10,000 troops bringing in lifting machinery, tents and water
puriﬁcation units. The press rushed to the scene, spreading word of the tragedy, ulti
mately leading to fund raising initiatives, both national and international. As more and
more donations poured into government collection and distribution centres the authorities
2
became unable to handle them, until ultimately the Chief Minister publicly announced
three days after the quake that “. . . no additional help of unsolicited kind is needed either
in the form of items or volunteers. . . We already have much, so much.”
For the longer term reconstruction eﬀort, bilateral donations from the United Nations
Development Program, the Asian Development Bank and a World Bank loan granted
the Indian Government US$358 million as a low interest emergency loan with a 30 year
payback period to add to its resources. In the third week of October 1993, the Govern
ment of Maharashtra invited NGOs for round table discussions of how to proceed with
reconstruction eﬀorts. The state government wanted to ensure that despite the large va
riety of agents involved in the reconstruction eﬀort, the houses met common standards of
earthquake safety. However, before the technological aspects of proposals could be fairly
assessed, the Indian media reported that politicians were attempting to exclude certain
designs: “. . . the chief minister and his cabinet colleagues are also believed to have taken
the psychological aspect into account. Nonconventional structures like geodesic domes
were ruled out because people are not used to living in them, it is learnt.” [Indian Express,
October 13th 1993]. By the time of the round table discussions, Adventist Development
and Relief Agency (ADRA) India had already contacted the villagers of Gubal about the
possibility of building geodesic domes as a safe housing device. The villagers agreed that
a sample geodesic dome could be built at the resettlement site for the villagers to inspect
and scrutinise. On the condition that the dome met village approval, the design would
be adopted as a rehabilitation measure. The Gubal village elder ultimately submitted a
request for 365 houses to be built, rather than the 182 originally planned.
2.2.2 Vigyan Ashram and engINdia
Early discussions of ADRA India with the people of Gubal concluded that safe housing
was of paramount importance in the villagers’ opinion. The horrifying experiences of
the Killari earthquake had turned much of the village into a graveyard, and none of the
surviving inhabitants wished to stay and live in old Gubal. After careful comparison of
earthquake resistant architectural designs, ADRA India decided upon the geodesic dome.
Vigyan Ashram, a research institution based in the village of Pabal had experimented with
the geodesic dome for a number of years. Dr Kalbag, Vĳay Kumar and other likeminded
engineers based there developed a modiﬁed version of the dome called the “Pabal” dome,
which was subsequently marketed by a number of entrepreneurial small scale industries,
producing prefabricated components as part of low cost doityourself kits and taking
orders to build and maintain the domes. Vigyan Ashram has continued to manufacture
and supply dome kits since the 1993 Killari earthquake, focusing on making lowcost, safe
and comfortable housing that is aﬀordable for the lowermiddle class of both urban and
rural populations.
As well as a research institution, Vigyan Ashram serves as a science centre for rural
3
youth. The centre was founded in 1983 by the late Dr Kalbag. Places are oﬀered both to
students in India’s public high school system, and also to school dropouts frustrated by
conventional education. Vigyan Ashram’s philosophy [Kalbag, 2004] places an emphasis of
teaching science through practice; to experiment, measure and record data, to recognise
patterns and to form and test hypotheses. The course is novel in two aspects: what
it achieves through student input and how it beneﬁts the village economy. Students
undertake projects in numerous areas including water resource development, construction,
workshop technology, energy, transport, environment agriculture, and home & health.
engINdia
1
is a partnership between 6 students from the University of Cambridge, Mas
sachusetts Institute of Technology (MIT) and the Indian Institute of Technology Bombay.
An expedition was conducted during the summer of 2005 to the area of Pabal, Maharash
tra, where the engINdia team worked with Vigyan Ashram and the local community to
gain an understanding and appreciation of the development issues concerning rural India
which could be tackled through engineering.
Through this relationship, MIT have established a FAB LAB in Pabal through their
“Bits and Atoms” program, with the intention of providing Vigyan Ashram and the local
community the necessary tools to empower them to solve their own engineering problems.
FAB LABs
2
share core capabilities, so that people and projects can be shared across
them. The kit includes: a computer controlled lasercutter, a large scale milling machine,
a signcutter, a micron resolution precision milling machine and programming tools for
lowcost, highspeed embedded processors.
2.2.3 The Water Bank project
The Water Bank project is a recent rural housing project intended to house 40 families
in Maharashtra. Work that began in January 2005 has been halted due to concerns over
the loading applied to a series of geodesic domes which form a large portion of the whole
complex. The domes have soil packed over them which was never anticipated in Vigyan
Ashram’s original design.
The complex is an experimental 8 million rupee investment (£116,000 at Jan 2009
rates) to house 40 families in Ankoli, Maharashtra. Of the 120 domes supplied to the
Water Bank housing project 40 are planned to be subterranean and are therefore aﬀected.
The complex is intended to provide each of the families with a guaranteed 2,000 litre water
supply yearround, a 300 square foot greenhouse, 350 square foot cave house, 900 square
foot work area, a terrace, a courtyard and 0.5 acres of land. The intention of the project is
to promote sustainable housing and fulﬁl the desires of the residents to be selfemployed.
1
http://www.engindia.net/
2
http://fab.cba.mit.edu/
4
2.2.4 Structural analysis
When the Water Bank project halted, Vigyan Ashram and engINdia approached the
student led charity Engineers Without Borders UK
3
to request that more research be
conducted into the design of the geodesic domes manufactured in Pabal.
There is a clear need for a method to model the structural response of the geodesic or
“Pabal” dome under various loading conditions. Vigyan Ashram’s original brief requested
that an assessment of the loading encountered by the domes should be performed and
design alterations should be recommended to accommodate the expected loading.
Following discussions with Vigyan Ashram, it emerged that they wished to alter the
original brief to focus the design of a bespoke structural analysis package. This would
allow Vigyan Ashram to analyse the structural response of the domes inhouse without
the need for buying costly ﬁnite element software licences or requesting outsider help.
This would be a further step to empower the locals to solve their own engineering prob
lems in the future. An inhouse analysis method would allow assessment of the dome’s
structural response in situations previously not considered. Vigyan Ashram hoped this
would increase the number of potential uses of the structure and its market demand.
If appropriate, students at Vigyan Ashram would also use the package to further their
understanding of structural behaviour.
2.3 Geodesic domes
2.3.1 Buckminster Fuller
Engineers and architects have always held a special interest for structural systems that
enable them to cover large spans with minimal interference from internal supports. It
is perhaps no surprise then that dome structures, capable of encompassing maximum
volume with minimum surface area, are one of the oldest structural forms and have been
used in architecture since the earliest times. The earliest geodesic dome was designed and
built in 1922 by Walter Bauersfeld in Jena, Germany [Makowski, 1979]. It was not until
the 1950s, and largely due to the eﬀorts of Buckminster Fuller, that the geodesic dome
became a vogue form of design.
Buckminster Fuller was an American born in Massachusetts, USA. Though he never
possessed formal qualiﬁcations in engineering or architecture, he inﬂuenced numerous
architects and some engineers to a greater extent than many eminent members of their
professions [Zung, 2001]. He ﬁled for a patent in 1951 for an improved version of the
geodesic dome design, which has since been used in structures such as the Tacoma Dome
(WA, USA), Poliedro de Caracas (Caracas, Venezuela) and The Eden Project (Cornwall,
UK).
3
http://www.ewbuk.org/
5
Buckminster Fuller’s passion for geodesic structures came from the design’s aﬃnity
with Nature, an eﬀect he described as the “energeticsynergetic” geometry of his domes.
“Energetic” refers to Fuller’s belief that Nature always builds the most economic struc
tures. He claimed that geodesic domes built upon principles embodying force distributions
similar to those of atoms, molecules, and crystals, would form the lightest, most eﬃcient
forms of construction. Fuller deﬁned synergy as the “integrated behaviours of nature and
the behaviour of a whole system unpredicted by the behaviour of any subassembly of
its components.” For example, hydrogen and oxygen gases, when combined in the right
conditions produce water, and this could not be predicted by the individual properties of
either gas. Similarly, the geodesic dome exhibits a stiﬀness and rigidity greater than that
predicted based on the sum of the individual components that make it.
2.3.2 Deﬁnition
Figure 1: Isosahedron Figure 2: Great circles [Morgan, 1981]
Buckminster Fuller based his original design on the sphere division of an icosahedron,
(see Figure 1) although geodesic domes have since been constructed using octahedron
and dodecahdreon symmetry systems to circumvent Buckminster Fuller’s patent. The
aforementioned shapes are all part of the family of platonic solids  polyhedra made up
entirely of congruent regular polygons. An icosahedron exploded onto the surface of a
sphere produces twenty equilateral spherical triangles, the vertices of which may also be
described by the intersection of three great circles (circles with a diameter equal to that of
the sphere) and are referred to as geodesic points, such as A, B and C in Figure 2. Fifteen
great circles also fully deﬁne the primary bracing of a geodesic dome. If the chords that
join the vertices are straight lines rather than curves, planar triangles are formed and this
creates the geodesic network commonly used in structures.
The primary bracing is truly geodesic, but impractical to use in most circumstances
as members quickly develop excessive slenderness ratios as the diameter of the dome
increases. To obtain a more regular network, a secondary bracing is introduced, modularly
dividing each equilateral triangle into a number of subdivisions. There are two possible
classes of geodesic subdivision; for Class I subdivision, the dividing lines are parallel to the
6
Figure 3: Class I subdivision [Motro, 1984] Figure 4: Class II subdivision [Motro, 1984]
edges of the primary bracing (Figure 3); in Class II, the dividing lines are perpendicular to
the primary bracing (Figure 4). Class I subdivision produces geometry where the edges of
the triangle lie on a great circle, which leads to simple design of a hemisphere with planar
connections; this may not be achieved with a Class II subdivision. Class II subdivisions
require a smaller number of bar lengths, which is advantageous for fabrication; however,
the diﬀerences between individual bar lengths are resultantly greater in a Class II dome,
and this produces a less uniform stress distribution. Additionally, Class II domes can only
be achieved with an even frequency of subdivision.
(a) V2 (b) V4 (c) V8
Figure 5: Example subdivision frequencies [Motro, 1984]
A subdivision, or “frequency” is deﬁned by the number of triangles each edge of the
primary bracing is divided into (Figure 5). The frequency is often referred to in short
hand as a number, with the preﬁx “V”. It should be noted that if secondary bracing is
introduced, the triangles are no longer perfectly equilateral  the bars forming the skeleton
show variations in length, and the number of diﬀerent lengths required to fabricate the
dome increase with the frequency of subdivision.
Odd order frequency domes cannot produce a hemispherical shape, as an equatorial
perimeter ring is only produced for even order frequencies. Odd order frequency subdivi
sions are nominally referred to with the suﬃx 3/8ths or 5/8ths, to indicate respectively
if the ground ring is above or below the equator of the geodesic sphere.
7
2.3.3 Explanation of behaviour
The manner in which a braced dome behaves depends on the conﬁguration of the members.
Fully triangulated domes, such as geodesic domes have a high stiﬀness in all directions and
are kinematically stable (no mechanisms) when idealised as a space truss. A dome which
is not fully triangulated is not kinematically stable when idealised as a truss and stiﬀnesses
may vary greatly in diﬀerent directions on the dome’s surface. While radial cable domes
may exhibit greater stiﬀness to uniform loads, triangulated domes demonstrate greater
stiﬀness to non uniform and concentrated loads [Kardysz et al., 2002].
The forces in a geodesic network are an equilibrated combination of tension and com
pression. Tension forces are global and continuous, while compression forces are local and
discontinuous. Buckminster Fuller coined the term tensegrity, a portmanteau of tensional
integrity, to convey the concept of coherence and resilient elasticity of geodesic networks.
Richter [1975] argues that the use of geodesic subdivision produces structures of greater
strength than conventional braced domes. The triangle is a planar ﬁgure which has
maximum rigidity accomplished with the least eﬀort. Symmetrical triangular systems
provide the most economic energy ﬂow, and a geodesic network produces a structural
form with self stabilizing properties.
2.3.4 Strengths and weaknesses
A dome is a typical example of a synclastic surface, where surfaces are of positive Gaus
sian curvature; i.e. where the curvature of any point is the same sign in all directions.
Synclastic surfaces are not developable; i.e. they can not be ﬂattened into a plane without
shrinking or stretching of the chords. This is why in practice domes cannot be built from
one uniform member size. However, Makowski [1965] points out that geodesic domes are
exceptionally good in this respect; even for sizable spans only a small number of diﬀer
ent bars sizes are required, making the structure ideal for optimising the manufacture
of components and prefabrication. Morgan [1981] describes how the Kaiser Aluminium
Company (USA) was one of the ﬁrst companies to take out a full license under Buckmin
ster Fuller’s patent for a concert hall in Honolulu, Hawaii. The 44.2m dome was erected
within 22 hours of the components arriving, and within 24 hours a concert was held at
the venue by the Hawaiian Symphony Orchestra to an audience of 1832.
In high risk earthquake areas such as Japan, quake resistant designs have evolved
through trial and error exercises over the centuries. Modern design guidelines have
emerged from such empirical data; the rules listed below are extracts from a planning
aid for architects and engineers [Dowrick, 1978] and are used to highlight some of the
properties that make the geodesic dome a highly earthquake resistant design.
• The degree of compactness in a building correlates with its resistance against seismic
shock. A hemispherical structure is the most geometrically possible compact form
8
of construction, enclosing maximum volume with minimum surface area.
• Reinforcing elements should be rigid and symmetrically organised, located close to
the building boundary and distributed evenly throughout the structure; the geodesic
skeleton does exactly this.
• The centre of gravity should be as low as possible. A dome has a lower mass point
than any cuboid structure of similar proportions.
• The construction should be elastic and deformable to a certain extent, especially
through the core grid. The geodesic dome exhibits this behaviour, and empirically
has been proven to survive even severe earthquakes with only minor cracking in the
cladding.
Geodesic domes produce extremely light skeleton structures that are very stiﬀ and rigid,
enclosing a large area without need for internal supports. Due to the light weight, the
round shape of the dome perimeter, and the generally uniform load distribution of geodesic
dome structures, deep foundations are not normally required. Construction of shallow
foundations allow a considerable saving in time and money over deep foundations.
The geodesic dome is, however, not without disadvantages. One complaint is that the
perimeter chords following the shape of an icosahedron produce an irregular or ragged line
that may be objectionable on architectural grounds; the aesthetic appearance of the dome
is largely dependent on how these closures are treated. The most common objections are
functionalist; due to the hemispherical shape of the structure, eﬀective sound isolation is
diﬃcult to achieve through partitioning of rooms, leading to a perceived loss of privacy.
Furniture, unless custom made, also presents a problem due to the curved walls of the
structure which can result in some loss of space.
2.3.5 The Pabal dome
Vigyan Ashram’s Pabal dome design is a developed version of the basic geodesic dome
structure. The Pabal dome is based on a V3 5/8ths Class I dome as shown in Figure 6.
The dome is made of prefabricated equal angle struts and disc hub joints made from
mild steel. A house kit may normally be assembled in one day using only simple nut and
bolt construction, without the need for specialised equipment. The kit requires no brick
masonry, so is easy to transport. The ﬁnished skeleton structure is clad using multiple
layers of chicken wire reinforcing mesh and ferrocementing (or “guniting”) technology; a
form of spray on concrete that is cheaper, lighter and faster than cladding using masonry.
The basic structure is a cost eﬀective form of design, costing only RS.200/ft
2
(£9.21/m
2
,
January 2009). Domes can be interconnected to form interesting housing designs and add
further rooms or space. The cladding forms an insulating thermal mass that keeps the
house cool all year round.
9
Foundation closure Disc joint
Door
Figure 6: Pabal dome
2.4 Braced dome analysis
2.4.1 Idealisations
Mullord [1984] states that before any engineering structure can be analysed, it has to be
represented by an idealised mathematical structure whose behaviour models suﬃciently
closely that of the original structure. Two idealised methods exist for braced dome struc
tures: equivalent shell methods and discrete structure methods. The equivalent shell
method may be divided further into two subgroups  orthotropic shell theory and ﬁnite
diﬀerence theory. Both methods use approximations which aim to smear the eﬀect of
discrete members uniformly over the surface of an equivalent shell, leading to a set of gov
erning diﬀerential equations that are solved using a harmonic solution. The second group
of methods allow the analyst to tackle the discrete structure directly. Again, this may
broadly be split into two subgroups
4
: space truss analysis, where all joints are assumed
pinned, and space frame analysis, where joints are assumed continuous. Both discrete
structure methods produce a large set of simultaneous equations that can only be feasibly
solved with the aid of a computer.
Equivalent shell methods are best suited to early design work or for structures too
large to be analysed discretely. Over the years, computers have become more aﬀordable
and more powerful, and the use of ﬁnite element analysis has become increasingly popular
method of modelling structures that are too time consuming or complex to resolve by hand
[Coates et al., 1988]. Vigyan Ashram favoured the development of a spreadsheet based dis
crete structure method, as such software is widely available and signiﬁcantly cheaper than
dedicated analysis packages. Matlab and C++ based FEA programs were alternatives
also considered, but were ruled out due to the required presupposition of programming
4
Other direct analysis methods do exist, for example, the equilibrium method discussed by Quintas and
Avila [1993] or Endogenous force analysis [Burkhardt, 2007], but space frame and space truss idealisations
are by far the most common.
10
knowledge in order to understand the workings of the structure stiﬀness method. It is
important to understand that Vigyan Ashram’s students will have limited programming
knowledge due to the vocationalstyle training used there to teach. Microsoft Excel 2003,
a widely available spreadsheet package that Vigyan Ashram are familiar with using, was
mutually decided to be the most appropriate platform to develop the solution with.
2.4.2 Analysis
Discrete structure methods can be modelled using linear or nonlinear analysis. Linear
analysis assumes a linear stress strain relationship, and together with information about
the material behaviour can be used to check for local member or joint failure. However,
instability eﬀects beyond the point at which the material yields require nonlinear analysis
to account for member eﬀects such as plastic yielding. Nonlinear analysis methods have
been extensively studied by many authors [Meek and Loganathan, 1989], and are based
on either the ﬁnite element method or on the beamcolumn approach.
To decide which analysis method was most appropriate, it was important to under
stand the level of complexity of the diﬀerent methods. Nonlinear methods are generally
more technical and more computationally exhaustive, but produce a more accurate so
lution. However, a more accurate, more technologically complex method was not neces
sarily the most relevant technology for Vigyan Ashram to use. Vigyan Ashram required
a straightforward analysis method that produces sensible design information for a com
prehensive range of loading options. Linear analysis, being less computationally intensive
and requiring less engineering knowledge to understand was therefore better suited for
use in Excel.
Figure 7: Brown University’s 2D FE program
11
2.4.3 Previous work
The concept of using Excel for the purpose of structural analysis is not entirely new. Teh
and Morgan [2005] highlight the beneﬁts of the platform for teaching the stiﬀness method
to ﬁnal year university students at Curtin University of Technology, Australia. Similarly,
a method of forming basic 2D pin jointed analysis developed by Brown University, USA
5
is available freeware for solving simple nodal applied loads (Figure 7).
However, these examples are limited to two dimensional analysis of a limited number
of members. The interfaces are also at a very basic skeleton level  requiring the user to
specify each nodal coordinate, member stiﬀnesses, ﬁxities and each nodal load, resulting
in a very time consuming and unnecessary data entry exercise. Some of these programs
also make use of built in matrix manipulation functions that are not designed to work
with large matrix problems. Vigyan Ashram’s request was for a user friendly interface
that would streamline the analysis process of a geodesic dome and work with a variety
of diﬀerent load situations. To achieve this, a fresh approach and a bespoke design were
required.
3 Theory
The following section introduces the underlying theory required to understand how the
FEA package covered in Section 6 was developed. The emphasis here is on the processes
involved in forming and solving a system, rather than justiﬁcation of why a particular
method is most suitable (this is covered in Section 2.4).
The reader is assumed to have a grasp of matrix mathematics; understanding matrix
addition, subtraction, multiplication, inversion and transposition
6
. A previous under
standing of the stiﬀness method will be helpful, but is not essential.
3.1 Stiﬀness method
The stiﬀness method allows us to analyse a structure which is an arbitrary assembly of
simple structural members by breaking down the components into “elements” (members)
and “nodes” (joints). It can deal with a wide range of design situations, including space
frame structures (where joints are treated as continuous). However, only the method for
a space truss system needs to be discussed here for the reasons described in Section 2.3.3.
The basis of the stiﬀness method for a structure limited to pin jointed elements is that
every member has analogous properties to those of a spring; that is, an axial load F may
be applied to the end of a spring, and this is internally carried, causing a resultant change
in length u.
5
http://www.engin.brown.edu/courses/en3/notes/Statics/Structure_tutorial/Structure_tutorial.htm
6
If not, most maths textbooks cover the topic of matrices; see for example Stroud [2001].
12
The relationship between applied load and the resulting displacement is a property of
the element, known as a spring constant, or stiﬀness k (Equation 1), expressed as
F = ku . (1)
It is possible to model the response of a structure by connecting together a system of
individual spring elements and solving to determine the displacements. In order to ﬁnd
a structure’s response to a known set of forces, a way of determining the stiﬀness of each
member must be found. Though the derivation of element stiﬀness is quite elementary,
it has been included to help explain the steps taken to post process the spreadsheet data
in Section 6.4.1.
Intuitively, if the section is made larger, or a diﬀerent material is used, the value of k
will change. Consider the expressions for stress (σ) and strain () in an element
σ =
F
A
(2)
and
=
u
L
, (3)
where F is the axial force in the member, A is the area of the cross section, u is a
resultant change in length and L is the length of the original member. Now, for a linear
elastic material, these expressions may be related using Hooke’s Law:
E =
σ
=
F/A
u/L
=
FL
uA
, (4)
and rearranging gives
F =
EA
L
u . (5)
Comparing Equation 5 to Equation 1, it is clear that an expression for the spring
constant in terms of geometric and material properties of the material exists; k =
EA
L
.
Now that the stiﬀness for an individual element has been determined, the principle
needs to be extended to work for a three dimensional system.
7
In 3D the expression
relating forces and displacements is much the same, however a total of six degrees of
freedom (DOF) must now be considered for each element, as each node can displace in
three dimensions: u, v and w (Figure 8).
Rather than dealing with multiple linear expressions, it is convenient to express the
local stiﬀness of an element using matrix notation:
7
The reader may wish to follow a worked example for a 1D or 2D structure stiﬀness method from a
textbook, such as Mullord [1984], before reading the remainder of this section.
13
Figure 8: DOF for a pin jointed element in a local coordinate system
_
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
_
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
_
f
x
1
f
y
1
f
z
1
f
x
2
f
y
2
f
z
2
_
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
_
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
_
=
_
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
_
EA
L
0 0 −
EA
L
0 0
0 0 0 0 0 0
0 0 0 0 0 0
−
EA
L
0 0
EA
L
0 0
0 0 0 0 0 0
0 0 0 0 0 0
_
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
_
_
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
_
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
_
u
1
v
1
w
1
u
2
v
2
w
2
_
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
_
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
_
(6)
or, in shorthand form
{f
local
} = [k
local
] {u} . (7)
Equation 7 uses a local, righthand rule coordinate system x’, y’, z’ (with the x’ axis
aligned with the direction of the truss element as shown in Figure 8). In a 3D structure,
the local Cartesian coordinate system is unlikely to coincide with the coordinate system
of the global axis X, Y, Z. In order to obtain meaningful stiﬀness matrix, a transformation
matrix [T] is applied to the local stiﬀness matrix to obtain a common a global stiﬀness
matrix for each element
8
.
[K
global
] = [T]
T
[k
local
][T] , (8)
where
[T] =
_
_
R
0
0
0 R
0
_
_
, (9)
and the general form of R
0
is
8
The derivation of this equation has been omitted, but is covered in Coates et al. [1988] and most
good texts on FEA.
14
[R
0
] =
_
¸
¸
¸
¸
¸
_
L
x
/L L
y
/L L
z
/L
−L
x
L
y
cos γ−LL
z
sin γ
L
√
L
2
x
+L
2
z
√
L
2
x
+L
2
z
cos γ
L
−L
y
L
z
cos γ+LL
x
sin γ
L
√
L
2
x
+L
2
z
L
x
L
y
sin γ−LL
z
cos γ
L
√
L
2
x
+L
2
z
−
√
L
2
x
+L
2
z
sin γ
L
L
y
L
z
sin γ+LL
x
cos γ
L
√
L
2
x
+L
2
z
_
¸
¸
¸
¸
¸
_
, (10)
where L is the length of the member, and its projected lengths on each of the global
axes X, Y and Z are L
x
, L
y
and L
z
respectively. The angle γ is a rotation of the element
about the x’ axis in Figure 8 to align the cross section of the element. In the case of a pin
jointed (truss) structure, the shape and orientation of the cross section is unimportant as
bending moments are ignored. Hence, for a space truss, γ = 0 and Equation 10 simpliﬁes
to
[R
0
] =
_
¸
¸
¸
¸
¸
_
L
x
/L L
y
/L L
z
/L
−L
x
L
y
L
√
L
2
x
+L
2
z
√
L
2
x
+L
2
z
L
−L
y
L
z
L
√
L
2
x
+L
2
z
−L
z
√
L
2
x
+L
2
z
0
L
x
√
L
2
x
+L
2
z
_
¸
¸
¸
¸
¸
_
. (11)
Once every element stiﬀness matrix has been transformed into a global coordinate
system, it only remains to assemble all the matrices to form the structure stiﬀness matrix.
1 2 3 4 5 6
1
2
3
4
5
1 2 3 4 5 6 7 8 9 10 11 12
6 1
2
4 5 6 7 8 9
3
4 4
5 5
6 6
7 7
8 8
9 9
10
1 2 3 10 11 12
11
1 12
2
3
10
11
12
K
elem1
6x6
K
elem1
6x6
K
elem1
[K]
6x6
(12x12)
K
elem1
K
elem2
K
elem3
K
elem3
K
elem3
K
elem3
Figure 9: Assembly of the structure stiﬀness matrix from element stiﬀness matrices.
In FE programs, this is often achieved through the use of a steering matrix which
links the local numbering scheme to the global numbering (see Figure 9). Nodes that
are shared between elements will have overlapping contributions to the same cells of the
stiﬀness matrix.
For a truss, each element stiﬀness matrix is 6 × 6 and the ﬁnal structure stiﬀness
matrix size [K] is determined by the number of nodes multiplied by the number of degrees
of freedom per node. For example, the structure stiﬀness matrix in Figure 9 would have
15
to be composed of four nodes, as 4 ×3DOF produces the 12 ×12 matrix shown.
bandwidth
bandwidth
(a) Low bandwidth
bandwidth
bandwidth
(b) High bandwidth
Figure 10: Structure stiﬀness matrix bandwidth
The node numbering system used in the structure is often critical to the eﬃciency of
the solver, especially for large structural problems with thousands of elements. An eﬃcient
numbering system will produce a narrowly banded set of values on the leading diagonal
of the populated matrix. The greatest width of any row in the matrix is termed the
“bandwidth”. A less eﬃcient numbering system produces a much wider spread of results
symmetrically about the leading diagonal and therefore produces a larger bandwidth (see
Figure 10).
Finally, a matrix version of Equation 1 has been developed, in terms of a known global
force vector {F} and the structure stiﬀness matrix [K]
{F} = [K] {U} . (12)
3.2 Solution methods
In order to gain a useful output from Equation 12, a method of solving it is required. First,
boundary conditions are applied to the problem to deﬁne it uniquely. This is normally
achieved through ﬁxing certain degrees of freedom in the structure to represent anchorage
to the ground.
Solving Equation 12 is not as straightforward as it may seem, as division is not possible
in matrix algebra. The closest equivalent matrix operation that emulates division is to
ﬁnd its inverse [K]
−1
, where
[K][K]
−1
= [I] , (13)
and [I] is an identity matrix, the matrix equivalent of unity. This allows us to then
solve Equation 12 as
16
{u} = [K]
−1
{F} . (14)
However, for a large matrix, the process of inversion is computationally intensive
and slow. Rather than tackle the inversion directly, considerably more computationally
eﬃcient methods are used. Trevelyan [2007] categorises solvers into two groups.
• Direct solvers. These solvers use methods such as Gaussian elimination and LU de
composition and are guarenteed to arrive at a solution if the matrix is nonsingular.
The number of ﬂoating point operations required to solve a system using a direct
solver increases with the cube of the number of equations, meaning that for large
matrix FE models, the solver quickly becomes a dominant part of the run time.
Direct solvers also suﬀer from an accumulation of roundoﬀ errors.
• Iterative solvers. These solvers, such as the conjugate gradient method, make
an initial guess and progressively iterate towards a solution until a convergence
tolerance is reached.
Many of these methods may be found in mathematical textbooks
9
, however only Gaus
sian methods will be mentioned in more detail here. Gaussian elimination has variants
such as GaussJordan elimination which has its own advantages and drawbacks.
The standard method, according to Kreyszig [2006], is a popular way of solving linear
systems, based on a systematic process of elimination in the rows of the matrix in order
to reduce the system to triangular form, as then the system may easily be solved by back
substitution. As Gaussian elimination is a particularly well documented way of solving
linear systems, it is unnecessary to reiterate the method in great detail here.
3.3 Post processing
Once the displacement vector in Equation 12 has been found using a method described
in Section 3.2, it is a relatively straightforward process to obtain the internal axial forces
that form an equilibrium with the externally applied force vector.
First, the transformation matrix from Equation 9 has to be applied to the global
displacement vector in order to revert the displacements to the local coordinate system:
{u
local
} = [T] {U} , (15)
The diﬀerence between the x’ directional nodal displacements u
2
− u
1
(see Figure 8)
give the overall change in length of the element which is then used to calculate the strains,
stresses and forces using Equations 2 and 3. A positive change in length indicates the
element is in tension, a negative change indicates compression.
9
See for example Kreyszig [2006].
17
3.4 Lateral earth pressure
The importance of the forces transmitted through the soil skeleton from particle to particle
was recognised in 1923 when Terzaghi presented the principle of eﬀective stress, which
applies only to fully saturated soils [Craig, 2004]. For a horizontal soil mass with the water
table well below the point of interest, the pore water pressure is zero, and hence eﬀective
and total vertical stresses at a depth below the surface, z, are equal and equivalent to the
weight of all material per unit area above that depth:
σ
v
= σ
v
= γ
sat
z , (16)
where σ
v
is the total normal (vertical) stress on the soil mass, σ
v
is the eﬀective normal
stress, γ
sat
is the bulk unit weight of the soil and z is the depth below some datum line
at the surface of the soil.
To relate the distribution of vertical stresses to those induced horizontally against a
retaining structure, Rankine’s theory of earth pressure is used, which considers the state
of stress in a soil mass when the condition of plastic equillibrium has been reached, i.e.
when shear failure is on the point of occurring throughout the mass.
According to Rankine, vertical stresses in the soil remain at a constant geostatic value,
wheras horizontal stresses increase or decrease depending on the local movements of the
soil. In an active pressure zone, the retaining wall is pushed away from the soil mass in
response to the action of the soil. In the passive zone, the retaining structure is being
pushed against a resisting soil mass.
Rankine found that the relationship between the horizontal and vertical stresses in
the active region
10
of the soil were related by an active pressure coeﬃcient K
a
, controlled
by the soil’s intrinsic shear strength parameter φ:
σ
h
= K
a
σ
v
, (17)
where
K
a
=
1 −sin φ
1 + sin φ
. (18)
4 Geodesic dome analysis
4.1 Introduction to Strand7
A commercial FEA package called Strand7
11
was used to build up an initial geodesic dome
model. Strand is a general purpose ﬁnite element analysis system, used in a wide range
10
The passive case calculation has been omitted for the reasons discussed in Section 4.3.3.
11
http://www.strand7.com/
18
of applications in mechanical, civil, structural, aeronautical and biomedical engineering.
The package was chosen because it was the most appropriate general purpose structural
analysis software available at Durham University’s School of Engineering.
While a powerful analysis tool, Strand would not itself be an appropriate piece of
software for Vigyan Ashram; licensing costs for this software are high, and it operates a
black box approach to analysis. The user is required to deﬁne a model and run the solver
without guidance or explanation. Strand will then produce a solution, assuming the model
is deﬁned correctly, without showing the intermediate calculation steps. It therefore is
not suited to teaching or explaining the FE method. However, to a competent user with a
good understanding of the theory behind the program, it is an eﬃcient tool for modelling
the behaviour of structures.
The following sections describe the generation of geodesic geometry in Strand, the
forces that were considered to act on the structure, and the results that it produced
under various analyses.
4.2 Coordinate generation
The mathematical determination of geodesic dome coordinates may be achieved in one of
two systems: Cartesian (presented by Davis [2007]) or polar (presented by Kenner [2003]).
Figure 11: Global coordinate system used for analysis
A polar coordinate system deﬁnes the position of a coordinate relative to the origin
at the centre of a sphere by a radius r and two angles θ and ψ, using algorithms based
on formex algebra (see SanchezAlvarez [1984]). It is a system particularly suited to
generating spherical geometries. However, a righthandrule Cartesian based coordinate
system was used for developing the Strand model, as the stiﬀness method discussed in
Section 3.1 is based upon Cartesian transformations. All forces and displacements in
FEA are expressed in Cartesian form, and element stiﬀnesses are described locally and
globally by Cartesian coordinates. It was hence logical that the basis of the model should
be deﬁned using a global Cartesian system (Figure 11).
19
A trial version of a package named CADRE Geo 6.0
12
was used to generate the geodesic
geometry for the Strand model. CADRE is a design utility for generating a wide variety
of geodesic and spherical 3D wire frame and surface models for Computer Aided Design
(CAD) and FEA applications. In addition to generating coordinate data, the package was
able to produce dfx ﬁles that could be imported directly into Strand, saving a considerable
amount of time on assembling the elements connecting nodes manually. Setting the radius
of a V3 5/8ths dome to unity in CADRE produced a set of coordinates which could be
scaled to deﬁne the geometry of the dome for any radius in a spreadsheet.
When designing a structure, the internal ﬂoor area produced may be of greater concern
than the overall external dimensions of the structure. This was indicated to be the case
for Vigyan Ashram, as the Pabal Dome’s assembly guide initiates the dome construction
by marking the desired dome radius at perimeter ring (ﬂoor) level. Normally, the radius
is deﬁned at the equator of a sphere, but, as discussed in Section 2.3.2, an odd order
frequency dome does not produce an equatorial perimeter ring because the structure is
not an exact hemisphere. CADRE presented an option to expand the perimeter ring of
the dome to equal the radius of the structure by applying a scaling factor to the coordi
nates (found to be 1.016). This factor was implemented as an option in the coordinate
generation spreadsheet.
4.3 Loading
This section discusses the diﬀerent loads that could act upon a geodesic dome structure,
explaining which were considered important and why. The method that nodal forces were
calculated from the applied loads is also described here. The totality of possible actions
on the structure were considered to be: self weight, wind, snow, soil, imposed, seismic
and impact.
4.3.1 Seismic and impact loads
It has already been demonstrated that the Pabal dome is an earthquake resistant form of
design (Section 2.3.4). Additionally, impact damage was considered very unlikely given the
residential use of the structure and that compound walls are often built around Indian
housing [Rolly, 2007]. As it was not therefore necessary to consider these actions for
normal design situations, it was assumed dynamic load eﬀects were beyond the scope of
this project and could be neglected from the analysis.
4.3.2 Snow
Snow loading is often critical to dome design [Knebel et al., 2002]. However, the Ma
harashtra region of India maintains a temperature of 2030
◦
C all year, and has a record
12
http://www.cadreanalytic.com/cadregeo.htm
20
coldest day of 10
◦
C [MapXL, 2006]. It was therefore safe to assume that snow loading
would not be a design issue and could also neglected from the analysis.
4.3.3 Soil
Datum Line
z z
σ
v
= γz σ
h
= K
A
γz
σ
v
σ
h
Z
X
Figure 12: Horizontal and vertical stresses acting on a subterranean dome (after Landry
[2002])
Soil loading was a necessary consideration because of the dome’s subterranean use in
the Water Bank project (see Section 2.2.3). It was assumed that the loaded dome would
act as a retaining structure against any soil piled against it.
In order to calculate the forces upon the structure, Rankine’s earth pressure theory
(covered in Section 3.4) was applied to calculate the total horizontal and vertical stresses
induced by the weight of the soil, based on a depth z below a user deﬁned datum soil
level, as shown in Figure 12. The surface area of the dome was then divided between the
nodes, allowing the calculated stresses to be converted into nodal forces.
The stresses were calculated for an active pressure case, with no resistive mass of soil
acting passively. It was conservatively assumed that the soil loading would be carried
by the dome framework alone, as this reduced the complexity of calculating horizontal
pressures for asymmetric loads.
The design assumes that the water table is below the foundations of the dome, which
was a reasonable assumption to make as the region is prone to drought [Infobase, 2000].
This meant that underground aquifers were likely to be relatively deep below ground level,
so that pore water pressure would not make a contribution to the stress state of the soil
as described in Section 3.4. It was assumed that even during the monsoon aquifers would
not recover to a level above the foundations.
The predominant surface deposit of the region surrounding Pabal village is Black
Cotton soil, as shown by the regional soil map in Figure 13 [Shroﬀ and Shah, 2003]. This
is formed from the subaqueous decomposition or in situ weathering of basalt rocks, which
21
Figure 13: Regional soil deposits of India (after Shroﬀ and Shah [2003])
produces the mineral Montmorillonite in an alkaline environment. Montmorillonite is a
dark swelling clay commonly known in India as Black Cotton soil, due to its prosperity
for growing cotton [Ranjan and Rao, 2000]. The load contribution due to the soil in the
analysis (Section 4.5) assumed that the dome was buried in Black Cotton soil. In later
work developing the Excel spreadsheet, more soil types were added as options for analysis.
4.3.4 Wind
Bhandari et al. [2003] give general guidance to the treatment of cyclonic storms and
hurricanes, which are characteristic of India, especially during the monsoon season. They
state that cyclonic storms do not normally extend further than 60km from the coast and
that hurricane actions normally occur in the north east of India. This places Pabal (which
is around 100km from the west coast) in a relatively low risk area.
The basic wind speed and design wind pressure were determined to Indian Standards
using IS875:Part 3 [Bhandari et al., 1987]. These were chosen in preference to Eurocodes as
they are more representative of Indian wind conditions (for example, they include factors
of adjustment that account for cyclonic wind speeds). The design wind pressure p
d
for
the Pabal region was determined to be 0.584kN/m
2
, the calculations and assumptions
behind this value may be found in Appendix C.
It was conservatively assumed that the wind pressure acts only from one direction and
positively. In actual fact, the external pressure proﬁle of a spherical structure is partly
negative (i.e. suction) for angles between approximately 45
◦
and 135
◦
relative to the
wind direction, which would act to reduce load on individual elements (see Figure 14).
Stabilising wind pressures produced at the back of the dome due to vorticies were ignored.
22
Figure 14: External wind pressure coeﬃcients for a spherical structure (from Table 18,
Bhandari et al. [1987])
For simplicity, forces on individual members were not considered; rather, forces were
considered to act on the structure as a whole.
In order to get the total wind load acting on a particular building, a force coeﬃcient
based on the geometry of the structure was used. Together with the eﬀective face area
A
e
and design wind pressure p
d
this gave
F
w
= C
f
A
e
p
d
. (19)
For a rough spherical structure, Table 20 of IS875 gave the force coeﬃcient as C
f
= 0.7.
The eﬀective area was taken to be projected area of the dome facing the wind, which in
turn was a function of the dome geometry. The total wind loading on the face of the dome
(found with Equation 19) was proportionally divided between the nodes by projected area
to obtain the forces on individual nodes.
4.3.5 Self weight and imposed loads
The main components that contribute to the weight of the structure are the elements
and the ferrocrete shell. These were initially represented using truss elements and plate
elements with speciﬁed densities in conjunction with Strand’s gravity function. This
automatically calculated the nodal force contribution when an analysis was run.
This arrangement was convenient for use in Strand, but was discarded in favour of a
spreadsheet based equivalent for two reasons; ﬁrst, to ensure the method of generating
forces would be common for Strand and Excel, so that results would be directly compara
ble; and second, because including plate elements in the analysis adds to the rigidity of the
dome. Though this may appear beneﬁcial, FE analysis of plate elements is a considerably
more complex procedure than the analysis of truss elements. It was desirable to keep the
analysis simple and focused on the geodesic skeleton. Aside from making the spreadsheet
stiﬀness solver easier to follow, this also allowed the dome to be considered in situations
23
where it may not be built with a reinforced concrete shell, for example, when providing
the framework for the greenhouses mentioned in the Water Bank project complex (see
Section 2.2.3).
In order to divide the self weight of the structure between the nodes in an equivalent
way to Strand’s gravity method, the area supported by each node had to be calculated.
The shell was assumed to be approximated by loaded triangular plates; hence each
triangle was bisected to evenly split the plate load distribution between the nodes. For the
V3 5/8ths dome, three unique chord lengths A, B and C make up the possible triangle
sizes for any given geometry. It was therefore possible to determine the area of each
triangle, and, as each triangle was supported equally by three corner nodes, a third of the
load on each plate was allocated to each node.
A similar, but simpler method was applied to ﬁnd the distribution of member forces
between nodes; assuming element homogeneity, then simply half the length of each con
necting member is supported by the node under inspection.
Figure 15: Example area division for a node
For example, in Figure 15, the node under inspection is bordered by ﬁve triangulated
sets of elements, each consisting of two element lengths; two of A and one of B. The area
of each plate in this case is
1
2
AB, and hence the total applied force on this node due to
the shell would be
F
shell
= 5 ×
1
3
×
_
(
1
2
AB) ×ρ
shell
g t
_
, (20)
where t is the thickness of the shell, ρ
shell
is its density and g is acceleration due to
gravity. All 5 members connected to the inspected node are of length A in this example,
hence the nodal applied force due to the elements’ self weight would be
F
elem
= 5 ×
1
2
×(A ×A
elem
) , (21)
where A
elem
is the cross sectional area of the members.
Once these equivalent support reactions for all nodes were calculated, the sum of
reactions in Strand for an isolated self weight load case was compared to those determined
24
in Excel to conﬁrm that they coincided.
More components of self weight, including mesh density and joint weight were later
included using the same distribution method to allow a greater amount of ﬂexibility
designing the structure. The imposed loading cases were similarly calculated.
4.4 Assumptions
In order to analyse the V3 5/8ths, Class I, Icosahedron based Pabal dome (with 61 nodes
and 165 elements), and based on some of the considerations discussed previously regarding
modelling forces on the structure, the following assumptions were made for the analysis.
• The good practice guidelines outlined in Eurocode 3, Section 5.1 apply.
1. Analysis shall be based upon calculation models of the structure that are ap
propriate for the limit state under consideration.
2. The calculation model and basic assumptions for the calculations shall reﬂect
the structural behaviour at the relevant limit state with appropriate accuracy
and reﬂect the anticipated type of behaviour of the cross sections, members,
joints and bearings.
3. The method of analysis shall be consistent with the design assumptions.
• The structure’s working life was assumed to be 50 years (category 4, EC030, Table
NA2.1), and the assumed building use was residential (category A).
• The door (see Fig. 6) and any windows in the dome structure were suﬃciently spaced
and installed in such a way that they would not reduce the structure’s rigidity.
• The analysis was not rate dependent, i.e. time related eﬀects on the structure, such
as creep, were ignored.
• The boundary conditions assume that the structure was fully ﬁxed at all its nodes
at ground level.
• Members were assumed to be their correct mathematical geodesic lengths, so that
no stresses were induced due to mismatched lengths.
• Any beneﬁcial contribution to the structure due to a stressed skin eﬀect from the
reinforced ferrocrete shell described in Section 2.3.5 was ignored.
• The components that make up the structure were homogenous and isotropic.
• The analysis was static (where conditions are independent of time, as opposed to
dynamic, which would consider behaviour under time varying conditions, such as
earthquakes).
25
• The structure was modelled as a space truss; assuming that all joints were pinned.
This was a reasonable assumption to make, for the reasons discussed in Section 2.3.3.
• The members were assumed to be suﬃciently squat so that buckling eﬀects could
be neglected.
• The analysis was isothermal (independent of temperature).
4.5 Analysis
The dome analysis was completed in Strand based on the assumptions listed in Section 4.4.
The purpose of the analysis was to explore which load cases were the most signiﬁcant and
to provide some solutions to which the Excel based FE method could be compared to for
accuracy.
The dome was set to a radius of 3m, with its base expanded to equal the radius (see
Section 4.2), as this is one of the most common kit conﬁgurations used at Vigyan Ashram.
Five load cases, described below, were considered individually and then in combi
nation, using the recommendations for safety factors supplied by the British Standards
Institution (BSI) Eurocode
13
0: Basis of Structural Design [BSI, 2007].
4.5.1 Load Case I: Self weight
Based on the technical drawings contained in Rolly [2007] and the information provided by
Vigyan Ashram, the thickness of the ferrocrete shell was estimated to be 50mm, bedded
on two grades of reinforcing mesh; a coarse (gauge19) square welded mesh and three
layers of ﬁner (gauge18) 1” chicken wire mesh
14
.
The joints and elements were assumed to be fabricated from mild steel. A nominal
mass of 1kg was speciﬁed for the joints, and the truss elements were assumed to be
25 × 25 × 3 equal angle sections made from rolled steel
15
; a choice reﬂecting the most
common member type used by Vigyan Ashram.
The properties of these structural components are summarised in Table 1.
4.5.2 Load Cases II & III: Soil
Two soil load cases were investigated (Figure 16). The ﬁrst assumed the soil would be
buried uniformly around the dome up to its tip, which for a dome with a radius of 3m is
a depth of 3.59m
16
. The second considered a case of nonuniform burial, a more realistic
situation where only half the dome was buried under a soil mass.
13
Eurocodes will henceforth be abbreviated to “EC” and will always refer to BSI [2007].
14
Ferrocrete associated data from Hartog [1984].
15
Steel data from Cobb [2007].
16
Recall that odd order frequency domes are not hemispherical (Section 2.3.2), and hence the height
of the dome is not equal to the radius.
26
Table 1: Element and joint properties
Material Parameter Symbol Value Unit
Steel unit weight γ
steel
77.2 kN/m
3
Young’s modulus E 200 GPa
Element cross section A
elem
157 mm
2
Ferrocement unit weight γ
shell
22 kN/m
3
Thickness of ferrocement t
shell
50 mm
Mass of coarse mesh m
c
1.15 kg/m
2
Mass of ﬁne mesh m
f
0.93 kg/m
2
Joint mass m
j
1 kg
(a) Load Case II: full burial (b) Load Case III: half burial
Figure 16: Soil loads considered
In both cases the soil used was assumed to be the regional topsoil: Black Cotton, with
the parameters
17
summarised in Table 2.
4.5.3 Load Case IV: Wind
The wind load case on the structure was calculated following the assumptions described
in Section 4.3.4. The prevailing wind direction was assumed to coincide with the global
X direction (Figure 11). Using Equation 19, the face wind force to be applied was
F
w
= 0.7 ×
_
5
8
π3
2
_
×0.584 = 7.22kN . (22)
17
Data based on the information provided by Ranjan and Rao [2000].
27
Table 2: Black Cotton soil properties
Parameter Symbol Value Unit
Shear angle φ 22 degrees
Bulk unit weight γ
bulk
20 kN/m
3
Active pressure coeﬀ. K
a
0.455
4.5.4 Load Case V: Imposed
A vertical imposed load of 0.6kN/m
2
was allowed for. This was a sensible value for roof
maintenance access, allowing for two workmen with tools to work on the roof.
4.5.5 Combination load cases
Three combination cases were considered, based on the safety factor recommendations for
permanent and temporary loads on structures set out in EC0.
1. Combination I: This was chosen to model a simpliﬁed case of full subterranean usage
of the dome. The full soil loading and self weight cases were factored as permanent
loads, and imposed loading was factored as a primary case temporary load. Wind
loading was ignored, as the dome would not be exposed to any wind.
Combination I = 1.35(Case I + Case II) + 1.05(Case V)
2. Combination II: This modelled a more realistic case of subterranean usage, where
the dome may be thought of as buried in part of an embankment. Imposed loading
was assumed to be the primary temporary load, wind on the exposed face of the
dome a secondary temporary load.
Combination II = 1.35(Case I + Case III) + 0.75(Case IV) + 1.05(Case V)
3. Combination III: This model reﬂected a more general usage of the dome structure.
The case considers the only permanent load to be the self weight of the structure.
Since the structure would be fully exposed to the elements, it was decided that wind
would be the primary temporary load. Workers accessing the roof would be a rare
occurrence, so imposed loading was factored as a secondary temporary load.
Combination III = 1.35(Case I) + 1.05(Case IV) + 0.75(Case V)
4.5.6 Results
The results of the Strand analysis are shown in Table 3. In the case of the maximum
displacements, u
s
, the global axis direction in which the displacement acted is speciﬁed
after the number.
28
Table 3: Strand analysis results
Case Max. Axial Forces (kN) Max. Disp. (mm)
Tensile Compressive u
s
Case I 4.2330 3.0366 0.8961 (Z)
Case II 10.3014 69.7234 6.8791 (Z)
Case III 33.3680 102.4470 7.8199 (X)
Case IV 1.0611 1.1811 0.2248 (X)
Case V 2.1608 1.5500 0.4571 (Z)
Combination I 21.8374 99.5681 9.9609 (Z)
Combination II 39.6312 143.9120 13.0223 (Y)
Combination III 7.7368 5.9753 1.5109 (Z)
4.5.7 Discussion
Several conclusions can be drawn from this basic analysis. Firstly, there is a clear indi
cation as to why the geodesic dome structures are failing in the Water Bank project; the
soil loading cases produce displacements an order of magnitude greater than for the other
cases. Similarly, the compressive forces in the members are considerably larger, which
indicates that premature failure of members and joints could be expected.
Under normal design conditions, the geodesic structures were found to deﬂect very
little. Even under Combination III, the factored deﬂections and axial forces remain low
for such a sizable structure.
An intuitive observation was that a symmetrical load (such as in Cases, I, II and V),
produced a symmetrical distribution of forces in the members. Perhaps not so intuitive
was that asymmetric loads produced much larger member forces. This can be seen clearly
from comparing Cases II and III: the former is buried with twice the volume of soil of the
latter, yet it is the asymmetric load case that develops considerably larger axial forces.
This supports the ﬁndings of Pakandam and Sarshar [1993], where a similar analysis was
carried out using snow loads.
It was also observed that the magnitude of the compressive forces exceed those of
tensile forces predominantly in the soil load cases. This makes logical sense, as large
external forces are acting to squash the dome inward, which would be expected to induce
larger compressive forces inside the geodesic network.
5 Joint testing
5.1 Introduction
This section of the report describes how a relationship between the discjoint properties
and the ultimate failure load of a joint was developed.
To validate the structural stability of a postanalysis dome, it was necessary to know
29
whether the axial forces calculated would lead to failure of the joint. The connections
used by Vigyan Ashram are discs rather than lap joints, which meant that the Eurocodes
for connection design did not oﬀer a method of estimating their capacity, as the code only
provides a relationship for joints with a constant cross section.
In order to develop a parameter linked relationship, a dimensional analysis was per
formed, identifying dimensionless groups of parameters that required further investigation
(the beneﬁts of this are discussed in Section 5.3). These parameters were then investigated
using both physical and computational testing:
• Some joint replicas were fabricated inhouse at Durham University and tested to
destruction to produce some “real life” data.
• Computational models of the discs were developed, testing a large number of con
ﬁgurations using stress analysis in Strand.
5.2 Pabal disc joints
Figure 17: Close up of Pabal dome joint
The joints used by Vigyan Ashram are stamped discs or “hubs” which have an array
of predrilled holes to which the truss elements of the structure are bolted to, as shown
in Figure 17. Two
18
diﬀerent types of joint are required to build any geodesic dome
conﬁguration  those that connect six elements (“hexagonal” discs) and those that connect
ﬁve elements (“pentagonal” discs).
18
Arguably fournoded connections are needed for the ground level perimeter ring, however these can
eﬀetively be produced by using only four of the six connections on a hexagonal disc.
30
The discs are fabricated from scrap mild steel, using a ﬂy press which stamps them
into the correct shape, including two distinctive crimps, which add bending strength to
the disc and also aid with stacking for storage and transport.
(a) Hexagonal disc (b) Pentagonal disc
Figure 18: Uniaxial loading of bolt array
Each individual bolt array is loaded uniaxially, as shown in Figure 18. For the most
common design conﬁgurations, tension forces are usually the prevalent axial force in the
structure (cf. Combination III in Table 3).
By inspection, a joint will always have a greater capacity against compressive failure
than tensile failure
19
, as the gross cross section of the material resists the load, whereas
in tension only the net area of the material resists the load (as reductions must be made
for bolt holes).
This means that even in cases where the largest magnitude axial force is compressive,
if evidence can be provided to support that a joint will not fail under that magnitude
of force in tension, then neither will it fail in compression. This simpliﬁes our task of
developing an equation to predict the failure load of the disc joints, as tensile failure
alone can be conservatively considered.
5.3 Dimensional analysis
Massey [1998] states that complete solutions to engineering problems can seldom be ob
tained using analytical methods alone, and experiments are usually necessary to determine
fully the way in which one variable depends on others. Dimensional analysis is a tech
nique that can be applied to reduce the number of experiments required to obtain these
relationships.
19
This is assuming that instability (buckling) eﬀects in compression may be ignored.
31
The ﬁrst task in the process of dimensional analysis was to decide which variables may
eﬀect the solution and express them in terms of their dimensional formula. No quantity
that may have aﬀected the problem was overlooked, save for quantities that only had
a indirect contribution
20
. The quantities considered were broken down into dimensional
components (mass [M], length [L] and time [T]) in Table 4.
Table 4: Dimensional analysis: step 1
Quantity Symbol Dimensional forumula
[M] [L] [T]
Failure strength f
u
1 1 2
Plate thickness t 0 1 0
Hub radius r 0 1 0
Axial force F
axial
1 1 2
The objective of dimensional analysis was to assemble the quantities into a smaller
number of dimensionless groups. To achieve this, the variables were arranged in a way
that removed their dependence on individual reference magnitudes, one after the other.
In this case, dependence on the mass dimension [M] was removed ﬁrst, by dividing
through by f
u
, giving Table 5.
Table 5: Dimensional analysis: step 2
[M] [L] [T]
f
u
/f
u
11=0 1(1)=0 2(2)0
t 0 1 0
r 0 1 0
F
axial
/f
u
(11)=0 1(1)=2 (2)(2)=0
As f
u
/f
u
gave unity, which is not a variable, it was eliminated. Dependence on the
length dimension [L] was then removed by dividing through by r, giving Table 6.
Table 6: Dimensional analysis: step 3
[M] [L] [T]
t/r 0 11=0 0
r/r 0 11=0 0
F
axial
/ (f
u
×r
2
) 0 2(1 ×2)=0 0
At this point, all dimensions had been eliminated from Table 6. r/r gives unity and
was hence eliminated, leaving two reduced dimensionless groups:
F
axial
f
u
r
2
and
t
r
(23)
20
In Table 4, Young’s modulus was neglected as this relates to the failure strength of the material.
Similarly, temperature eﬀects were ignored as these were material dependent.
32
In order to determine the relationship between these two groups, some experimental
work was required.
5.4 Tension testing
5.4.1 Joint fabrication
Vigyan Ashram initially planned to send some sample joints for testing at Durham Uni
versity; this was however ruled out due to expensive shipping costs. It was therefore
necessary to design and fabricate some replica joints. Engineering drawings of the joints
were subsequently drafted (see Appendix D), based upon the technical illustrations pro
vided by Rolly [2007] and advice of Vigyan Ashram.
Durham University did not possess a press suﬃciently powerful to create the bespoke
joints from mild steel, hence a custom built press was instead used to fabricate joints
from sheets of aluminium. Though not an ideal representation, the results would still be
suitable for comparison between practical and computational models.
The joint fabrication was a labour intensive process, and hence only a limited number
of discs were made. A total of six discs were made: two pairs of 1.5mm hexagonal and
pentagonal discs and one pair of 2.0mm hexagonal and pentagonal discs.
5.4.2 Apparatus
Oil pump control
Upper jaw
Lower jaw
50mm
Fine‐tune
oil strain
Position
dial
Oil gauge
Figure 19: Denison loading machine
A Denison tension/compression loading machine was used for the tests (Figure 19).
This machine was capable of applying a uniaxial tension force of up to 25kN, measuring
force/displacment data electronically into a Windows based software package. At the
time of use the machine was calibrated to a level complying with the United Kingdom
Accreditation Standard (UKAS).
33
Finding a method of clamping the discs that would accurately represent the conditions
to which the joint experiences in situ was problematic. A geodesic tensegrity network of
the dome is three dimensional, whereas the Denison could only emulate one dimensional
loading.
(a) Hexagonal rig (b) Pentagonal rig
Figure 20: Test rigs for joints
Several clamping methods were considered for the disc that would preserve the uniaxial
tension case, but were hindered by the 50mm clearance between the teeth of the top and
bottom clamps. It was decided to opt for a simple test rig using thumbscrews to attach the
joint to predrilled (5mm thick) steel bars considerably stiﬀer than the test joint, as shown
in Figure 20. This setup introduced a bending component into the tests, an undesirable
eﬀect, but was the most feasible option to allow experimentation to proceed.
5.4.3 Risk assessment
The experiment procedure was considered fairly low risk and only standard operating
precautions were employed. As aluminium is a ductile metal, nonexplosive failures were
expected. The Risk Assessment and COSSH form for the experimental work may be
found in Appendix F.
5.4.4 Method statement
The following operating procedure was used:
1. The Denison was switched on, and the oil feed activated to around half load, which
began to separate the top and bottom clamps hydraulically. The oil feed rate was
reduced to hold the Dension stationary when the top and bottom clamps were spaced
approximately at the right height to ﬁt in the test rig.
2. A joint was then selected for testing, and its thickness checked and noted at ﬁve
approximately even spaces around its perimeter using a micrometer (accurate to
34
±0.005mm).
3. The joint was attached to the test rig using thumbscrew connections and clamped
in the top and bottom jaws of the Denison.
4. The Denison datalogging software was loaded and the current position and registered
force were set as datum values.
5. While continuing to hold the clamp closed with the winding key
21
, the oil ﬂow was
increased so that the teeth continued to separate slowly. A load of approximately
0.2kN was applied before releasing the clamp.
6. The oil ﬂow was continually adjusted to keep the rate of separation at approximately
0.025mm/s throughout the test to maintain a smooth loading curve.
7. When the disc failed, the test was stopped, and the positions of the top and bottom
clamps were reset to an appropriate distance to repeat the above procedure for the
next joint.
5.4.5 Results
(a)
(b)
(c)
(a)
(b)
(c)
Figure 21: Tension test results
21
This was because the clamps are held together by the friction of the teeth. When little or no load
was applied, the teeth would slip free unless physically restrained from doing so.
35
The ﬁrst pair of 1.5mm thick joints were tested as a trial run, in order to gain familiarity
with the machine and experimental procedure before the full tests were run. The force
displacement plots for the remaining four joints are shown Figure 21.
5.4.6 Discussion
Several observations may be made about Figure 21. Each test plot begins approximately
linearly, save for a kink at point (a), corresponding to the point where the disc deforms
to align with the axial tension. Each disc reaches a peak capacity (b) after which the
joint begins to tear in shear. Slight recoveries in strength were observed at point (c) as
the joint crimps formed a necking region for torn material which temporarily impedes the
shearing action.
Of particular interest was that the pentagonal joints in each case failed at a lower
force than the hexagonal equivalents. This is counter intuitive as Figure 18 clearly shows
that for a common load, the pentagonal bolt array will have a greater equivalent area to
resist loading, so failure would be expected to occur hexagonal disc ﬁrst.
This diﬀerence may be due to the less than ideal simpliﬁcations made for this test.
In particular, the Tshaped clamping arrangement (Figure 20(b)) for the pentagonal disc
was fairly far removed from the uniaxial tension case that was of interest and this may
have adversely inﬂuenced its load capacity.
5.5 Stress analysis
A series of nonlinear joint analyses were run using Strand to estimate the force at which
joints would fail for various combinations of material parameters. Each test applied a
series of incremental loads to the plate, calculating the resulting stresses and displacements
in the material.
5.5.1 Model development
CAD models of the joints were created in Solidworks and imported to Strand. Problem
symmetry was taken advantage of to reduce the analysis to the loaded area of a single
bolt array (as shown by the hatched area of Figure 18).
Boundary conditions were prescribed along each of the two symmetry edges of the
element, preventing any translation or rotation, save for radial slip. An incremental force
was applied in the plane of the disc using a face pressure at each of the two bolt holes,
representing the axial force transferred to the joint through the connection bolts.
The plates were meshed for analysis in two stages: a tet6 element surface mesh,
followed by a tet10 brick mesh, both at a 3% mesh density. Before the investigation
commenced, the mesh quality was checked using a convergence test. Three progressively
ﬁner meshes were run to calculate the planar displacement of an element at the center of
36
(a) 3% mesh (b) 2% mesh (c) 1% mesh
Figure 22: Mesh convergence
Table 7: Convergence test results
Mesh grade Run time(s) Displacement (mm)
3% mesh 6 0.0183190
2% mesh 18 0.0186015
1% mesh 718 0.0187712
the outer bolt hole (Figure 22). The results (Table 7) conﬁrmed that the mesh quality
was reasonable, as the results were clearly converging on a solution at ≈ 0.0188mm.
Three diﬀerent plate thicknesses (1.5mm, 2.0mm and 3.0mm) and three diﬀerent mate
rial properties (6082:T6 aluminium, S275 steel and S450 steel) were investigated. Results
were repeated for both hexagonal and pentagonal material conﬁgurations, giving a total
of 18 separate analyses. Ideally, the disc radius should also have been varied to fully in
vestigate the dimensionless groups in Equation 23. However, a total of 54 analyses would
have been required to investigate 3 radii variations, an unfeasible amount of work given
the time frame of this project
22
. For duration of the computational work, the disc radius
was ﬁxed to 77.5mm.
(a) Aluminium (AL6082) (b) Steel (S275) (c) Steel (S450)
Figure 23: Sketches of material stressstrain relationships
A linear static analysis could not be used to predict the failure load, as this assumes the
material always behaves linear elastically. Nonlinear analysis (Section 2.4.2) was instead
22
The total required number of analyses to investigate a relationship increases to the power of the
number of variables.
37
used, as this models the behaviour of the material as parts begin to deform plastically,
until a limit is reached where deformation will continue indeﬁnitely with no increase in
applied force.
The model also assumed ﬁnite deformations, where the model progressively considers
the deformed shape from the previous load increment. This is more representative of
reality, but more computationally intensive than assuming inﬁnitesimal strains and small
displacements (where deformations are considered only relative to the original geometry).
Nonlinear analysis required a stressstrain curve to be assigned to each material to
inform Strand how the material would behave beyond yield. Material stress strain re
latioships are often complex (see curve (1) in Figure 23), so to simplify the analysis, a
perfectly plastic relationship was assumed, where ultimate failure was assumed at yield
(curve (2) in Figure 23). This was a conservative assumption to make, as all the materials
considered undergo plastic work hardening eﬀects that improve capacity before ultimate
failure.
5.5.2 Results
Figure 24: Plot of relationship between dimensionless variables
The forcedisplacement plots may be found in Appendix B. For each analysis, the axial
force at which the displacement appeared to tend towards inﬁnity (i.e. the ultimate load)
was estimated from the graph (using the weaker 6noded joint) and the dimensionless
variable values from Equation 23 were calculated. The data used be found in Table 8 and
the resulting graph from plotting the dimensionless groups is shown in Figure 24.
38
Table 8: Nonlinear joint analysis results
Material F
axial
(kN) t(mm) f
u
(MPa) t/r F
axial
/f
u
r
2
AL6082 6.7 1.5 205 0.0194 0.005441
AL6082 7.3 2.0 205 0.0258 0.005929
AL6082 8.0 3.0 205 0.0387 0.006497
S275 14.0 1.5 430 0.0194 0.005421
S275 15.1 2.0 430 0.0258 0.005847
S275 16.7 3.0 430 0.0387 0.006466
S450 17.7 1.5 550 0.0194 0.005421
S450 19.2 2.0 550 0.0258 0.005847
S450 21.2 3.0 550 0.0387 0.006466
5.5.3 Discussion
Figure 24 indicates some form of linear relationship may be established between the
two variables. It is unreasonable to include the radius in the expression calculated, as its
inﬂuence on the relationship was not rigorously tested. The linear polynomial of Figure 24
can therefore be simpliﬁed by ﬁxing the considered radius size as a constant,
F
axial
f
u
×77.5
2
= 0.0533
t
(77.5)
0.0044 , (24)
and rearranging,
F
axial
= f
u
(4.13t + 26.43) . (25)
The forcedisplacement curves (Appendix B) indicated that for the theoretical stress
analysis, the diﬀerence in capacity between hexagonal and pentagonal joint arrays was
minimal. Intuitively, the pentagonal joint capacity was always the greater of the two in
each test, in contrast to the practical test ﬁndings discussed in Section 5.4.6.
Table 9: Comparison of theoretical and practical joint capacities
Material f
u
(MPa) t(mm) r (mm) Lab F
axial
(kN) Theoretical F
axial
(kN)
AL6082 205 1.5 77.5 6.6 6.7
AL6082 205 2.0 77.5 9.0 7.1
As the pentagonal joint results were considered unreliable for the physical tests, the
tensile failure values for the hexagonal value, taken from Figure 21, were compared to to
the theoretical equivalents, calculated using Equation 25. The results shown in Table 9,
though not comprehensive, indicate that for low joint thicknesses the theory accurately
predicts the failure force, while for greater thicknesses it provides a 20% underestimation
of their capacity.
There are a number of reasons that the theoretical estimations become overly conser
vative.
39
A possible cause is that isotropic work hardening eﬀects were neglected in the the
oretical model (see Figure 23), and this contribution became more pronounced as the
thickness of the joint increased.
A tradeoﬀ between computational eﬃciency and accuracy had to be made, and it is
possible that the FE mesh was not ﬁne enough to obtain good results. Additionally, the
boundary conditions assumed in the model represented a worst case, and in reality limited
displacements take place at the assumed boundaries of the disc segement modelled.
A ﬁnal possible reason for the reserve strength of the joint is that the loaded segment
area was conservatively assumed. When the joint as a whole is loaded, it reaches a state of
equilibrium. As not all the loads on the joint may be of the same magnitude, it logically
follows that larger bolt arrays may be supported by a larger area of the joint than that
assumed, and therefore have a greater capacity of resistance.
6 Spreadsheet development
This section of the report describes how the FEA based package for analysing geodesic
domes works, with reference made to the theory of Section 3. The features that make
the layout user friendly and accessible, as well as limitations of the package, are also
discussed.
6.1 Package structure
The package was broken down into three Excel ﬁles, as a single spreadsheet became un
wieldy to run due to the large amount of data being processed. These ﬁles guide the user
through the stages of performing a dome analysis and design. First, the user deﬁnes a ge
ometry and loading case on the structure in “FEA_Input”. The spreadsheet subsequently
calculates the force vector which is read into the second spreadsheet “FEA_Solver”, where
the structure stiﬀness matrix is formed. Equation 12 is then solved for the structure
displacements. The ﬁnal spreadsheet, “FEA_Output”, postprocesses the displacement
results to give member forces in the structure. The worst case force produced is utilised
in a series of calculations that allow the user to optimise the structure to carry the load
and make savings on materials.
Start FEA_Input FEA_Solver FEA_Output End
Define
geometry,
materials and
forces
Force
vector {F}
Displacement
vector {U}
Axial member
forces, design
calculations
Figure 25: Flow chart of design process
40
6.2 FEA_Input
FEA_Input is the spreadsheet in which the user deﬁnes all the parameters of a particular
dome  geometry, materials, loads and partial safety factors. The main output of the
spreadsheet is the force vector.
The spreadsheet consists of ﬁve worksheets, described below.
• Summary: this is the only worksheet in the FEA_Input ﬁle requiring user inter
action. All the variables are deﬁned here, and the various worksheets tabs link to
these parameters to calculate forces. The force vector is calculated based on the
summation of various factored loads.
• Graphs: the level of soil applied to the structure above the base of the foundations
is calculated and represented graphically on this worksheet. Additionally, this page
shows which portion of the dome is considered loaded.
• Soil: the nodal forces in the structure due to soil loading are found using Rankine’s
theory for earth retaining structures as described in Section 4.3.3. Various soil types
may be selected from, imposing diﬀerent parameters on the calculations. The user
can also specify the depth and area of the dome that is loaded.
• Wind: this worksheet calculates the worst case eﬀect of a wind pressure acting on
one face of the structure, based on a userdeﬁned wind load (see Section 4.3.4 for
details).
• SW & Imposed: the nodal forces due to the self weight of the structure and due
to any speciﬁed imposed loading are calculated in this worksheet. The methodology
of this was previously discussed in Section 4.3.5.
6.3 FEA_Solver
The FEA_Solver spreadsheet emulates the stiﬀness method discussed in Section 3.1 to
form Equation 12. A Gaussian elimination method is then used to solve for the displace
ment vector {U}. The solver spreadsheet runs automatically when opened, with no user
input required.
6.3.1 Stiﬀness matrix assembly
The spreadsheet ﬁrst reads in the force vector and speciﬁed parameters from FEA_Input
and uses the information about the geometry to scale the coordinates in the “nodal
coordinates” worksheet. This in turn is used to calculate the total length L of each member
41
Figure 26: Calculation of project member lengths
and the axis projected lengths L
x
, L
y
, L
z
(as shown in Figure 26), using predetermined
information about how elements and nodes link together in the structure
23
.
[T]
T
[T] (Eq
n
s 9 and 11) [k
local
] (Eq
n
6)
[k
global
] (Eq
n
8)
Figure 27: Assembly of the global stiﬀness matrix
Each local element stiﬀness matrix [k
local
] is calculated using Equation 6. Using the
projected member lengths, [R
0
] and resultantly [T] are found for each element using
Equations 11 and 9 respectively, which may be transposed to ﬁnd [T]
T
. Finally, Equation 8
is used to determine the global element stiﬀness matrix (see Figure 27).
The numbering scheme used for degrees of freedom may appear abstract at ﬁrst, but
the system was chosen to add clarity to which degree of freedom was under scrutiny  the
last number is always “1”, “2” or “3”, referring to whether it was an X, Y or Z directional
DOF in the global coordinate system. The remainder of the number (disregarding the
last unit) describes which node the degree of freedom belongs to. For example, “312”
would refer to node 31’s Y directional DOF.
In order to assemble the full structure stiﬀness matrix from the 165 global element
stiﬀness matrices without resorting to a programming script, several intermediate stages
were used in Excel to get the data in a form that allows the summation of cells with
common values, as described below.
23
The knowledge of how the elements link up was a feature builtin to the spreadsheet, based on the
element and node numbering scheme used by Strand7.
42
MMULT
...
...
Figure 28: Intermediate matrix assembly step
1. A 6 × 183 Boolean matrix
24
[B] was created for each element. This matrix was
labelled such that only the rows with DOF associated with the local stiﬀness matrix
were considered, whereas the full list of 183 degrees of freedom were included for
the columns. Excel checked each cell’s row and column reference, producing a zero
if they do not match, and a one if they do (Figure 28).
2. For each element, the global stiﬀness matrix and the Boolean matrix were then
multiplied together using Excel’s built in matrix multiplication function “MMULT”.
This produces a partly assembled version of the structure stiﬀness matrix, which
shall be henceforth referred to as an intermediate [C] matrix. This can be thought
of as fully assembled in terms of columns but not rows (Figure 28),
[C] = [k
global
][B] . (26)
3. Once a [C] matrix was produced for each element, it was possible to sum all the
matching rows in the assembly to complete the full 183 × 183 structure stiﬀness
matrix. An Excel function known as “SUMPRODUCT” was used to achieve this:
for each cell in the 183×183 structure stiﬀness matrix, the calculation seeks the row
and column DOF reference of the cell and then looked for each corresponding row
and column match in the intermediate matrices. It then summed together the cell
values for all the matches and outputs the result in the structure stiﬀness matrix.
Figure 29 shows an example of the method for a simple set of matrices. For instance,
if the value of the assembled structure stiﬀness matrix circled in red was of interest, the
SUMPRODUCT function would seek its reference position (2,2) and look in the above
24
A Boolean matrix is a matrix with entries from the Boolean domain B = {0, 1}. Such a matrix can
be used to represent a binary relation between a pair of ﬁnite sets [Flegg, 1965].
43
Figure 29: Structure stiﬀness matrix assembly using SUMPRODUCT
transition matrices for the same reference numbers. In this case, it would ﬁnd two matches,
circled black. The matching cells are summed to get the ﬁnal value which is output in
the assembled structure stiﬀness matrix. The same principles apply for the calculation of
each stiﬀness value in the structure stiﬀness matrix.
6.3.2 Solver
At this point, the spreadsheet has fully deﬁned the force vector {F} and structure stiﬀness
matrix [K] in Equation 12. In order to solve the system, Excel uses a builtin macro
written in Visual Basic for Applications (VBA) to perform Gaussian elimination. This
solver method was chosen as it is popular and well documented, making it easier for
someone wishing to learn the mechanics to follow (see Section 3.2).
This code was adapted from one presented in Billo [2007], operating as a user deﬁned
array function
25
in the form: GaussElim(coeﬀ_matrix, const_vector), where coeﬀ_matrix
is set to equal the cell reference area of the structure stiﬀness matrix, and const_vector
is set as the cell reference area of the force vector; both having been ﬁrst reduced by
applying the perimeter ring boundary conditions.
25
An array function is a special method of calculation employed when working with arrays, matrices
and vectors. Rather than applying one formula to one cell, a formula is applied to an array of cells. For
more information, see Pearson [2008].
44
6.4 FEA_Output
The main purpose of the FEA_Output spreadsheet is to determine the internal axial
member forces by post processing the results of the structure stiﬀness method, and to
allow the user to tailor the fabrication details of the dome so that a safe and economical
design is produced.
6.4.1 Postprocessing
Figure 30: Calculation of member axial forces
The ﬁrst step taken by Excel is to import the displacement vector previously calcu
lated in FEA_Solver. Information about the member lengths and properties are also
read into the worksheet. The displacements are sorted by DOF for each element, and
the transformation matrix (Equation 11) is applied in reverse for each node to get the
displacements back in terms of the elements’ local coordinate system. Note that the
spreadsheet uses [R
0
] to transform each node’s 3 degrees of freedom, rather than applying
the full transformation matrix to the full force vector as described in Equation 9. The
methods are entirely equivalent, but the chosen approach is slightly more compact and
therefore a better method of displaying the information.
The local x’ (Figure 8) axis displacements are subtracted from one another, giving the
change in length of the element due to the applied external forces on the structure. The
strain in each member is calculated using Equation 3, and, by virtue of Hooke’s law, the
stress can be found using Equation 4. Finally, the axial force in each element is found
using Equation 2 (see Figure 30).
6.4.2 Calculations
The calculation worksheet was designed to produce a comprehensive document summaris
ing any analysis and design case. The layout was thereby set out in a printer friendly
format, so that pdf or paper based printouts could be made for permanent record. An
example calculation print out may be found in Appendix E.
The worksheet was designed, where possible, to the guidelines of Eurocode 3 [BSI,
2007]. The ultimate design load of the structure was set as largest magnitude axial force
45
calculated in the elements during post processing
26
. The spreadsheet makes checks for
safety for the Ultimate Limit State (ULS) and Serviceability Limit State (SLS).
The following four design checks were considered:
• Joint shearing failure (ULS): this check used the empirical relationship devel
oped in Section 5 to estimate joint capacity. Equation 25 is additionally factored
by a material constant to represent variable material strengths in steel.
• Bolt connections (ULS): the bolts were checked using guidelines laid out for steel
bolted connections in EC3, Section 3: Connections made with bolts, rivets or pins.
Checks were made for both shear and bearing resistance of the bolts.
• Angle section (ULS): this checked the capacity of the equal angle truss elements
used by Vigyan Ashram against failure in tension at their smallest net crosssection,
i.e. where holes are drilled for the connections, using the guidelines in EC380,
Section 3.10.3. Bending moment checks for the section were not required for the
reasons discussed in Section 2.3.3.
• Deﬂection limit (SLS): this check was introduced to avoid severe deformations
of the structure that would impede its appearance or functionality under working
loads. The deﬂection limit was set to L/360, which the Eurocodes presented as
the most suitable limiting value for a member clad in a brittle material subject to
cracking (EC359, Section 7.2).
The worksheet also calculates relevant fabrication details, such as the number of joints
and diﬀerent member lengths required to assemble the dome and utilises conditional
picture formatting to produce subdivision assembly diagrams for the dome (from Landry
[2002]) depending on which conﬁguration is selected. It should be noted that while the
fabrication and assembly details can be calculated for many diﬀerent dome conﬁgurations,
the FE solver method currently only considers the V3 5/8ths dome (see Section 4.4). The
expanded version of the fabrication details were created in anticipation that further work
may be carried out on the package to expand its analysis capability to any type of dome.
At present, the worksheet has warning features built in to inform the user if they attempt
a design mismatched from the analysis capabilities.
6.5 Discussion
Before the accuracy of the spreadsheet based analysis is compared to the benchmark
results set by Strand7, it is worthwhile discussing some of the features that make the
spreadsheet package userfriendly and well suited to the aims of this project.
26
The reason this assumption could be made without regard to whether it is tensile or compressive was
covered in Section 5.2.
46
Firstly, the stepbystep layout of the spreadsheet (Fig. 25) leads the user through
distinct stages of modelling, analysis and design. By keeping to this format, the layout of
the spreadsheets was orientated towards explaining the purpose of each step.
While it is entirely possible to perform an analysis using this package without in
depth understanding of the inner workings of the calculations, these were left visible
and accessible to the users interested in learning how the analysis works. To aid with
understanding the FE method, a userguide was produced to accompany the package,
explaining the basis of the FE method and how the spreadsheet emulates it. Black box
steps were avoided wherever possible in the spreadsheets. For example, the structure
stiﬀness matrix may have been more eﬃciently assembled using a VBA macro, but the
stepbystep approach implemented in the spreadsheet was felt to be more suitable, as it
demonstrates the process with greater clarity. The only exception to this was Gaussian
elimination, which had to be performed as a Visual Basic script due to the looping nature
of solving a problem using this method.
Use was made of Excel based features such as conditional formatting to highlight cells
of interest (e.g. passed or failed calculation checks) or to add conditional warnings if the
user attempts something inadvisable (e.g. spacing bolts closer than the recommended
limit). Additional advice was provided to the user using Excel’s cell commenting feature,
where hovering over a commented cell brings up an information box with additional
dialogue explaining its function.
Table 10: Spreadsheet based FEA results
Case Max. Axial Forces (kN) Max. Disp (mm) % Relative error
Tensile Compressive u
e
(cf. Table 3)
Case I 4.2329 3.0366 0.8960 (Z) 9.998 ×10
−5
Case II 10.2997 69.7237 6.8789 (Z) 2.907 ×10
−5
Case III 33.3654 102.4467 7.8202 (X) 3.836 ×10
−5
Case IV 1.0610 1.1811 0.2248 (X) 0.000 ×10
−5
Case V 2.1607 1.5500 0.4570 (Z) 21.188 ×10
−5
Combination I 21.8353 99.5682 9.9605 (Z) 4.016 ×10
−5
Combination II 39.6277 143.9114 13.0222 (Y) 0.768 ×10
−5
Combination III 7.7372 5.9747 1.5108 (Z) 6.619 ×10
−5
The results of the spreadsheet FE method are shown in Table 10. The boundary
conditions, forces and design assumptions made were identical to the case described in
Section 4.5. The relative error for each load case was found by normalising the maximum
Excel displacement u
e
by the maximum Strand7 displacement u
s
(from Table 3) and
expressing it as a percentage:
error =
_
1 −
u
e
u
s
_
×100 . (27)
47
For each load case, the relative error between Strand7 and Excel was found to be
negligibly small, indicating that the spreadsheet based FE method accurately imitates the
commercial package. The computation time to reach a solution was signiﬁcantly longer in
Excel (typically 3060s in the spreadsheet cf. to 3s in Strand), as may be expected, but as
the point of the exercise was not one of solver eﬃciency, this fact is perhaps unimportant.
7 Conclusions
Returning to the objectives deﬁned in the project plan, it is possible to assess how com
pletely the project objectives have been met. The main goal of this project worked towards
producing a Microsoft Excel based structural analysis method, to give Vigyan Ashram the
inhouse capability to perform structural analysis. This has been successfully met, with a
spreadsheet based system that reproduces the results of an considerably more expensive
commercial ﬁnite element package using the structure stiﬀness method. A wide variety
of diﬀerent load types and geometries can be applied to the dome, making the package a
versatile analysis tool.
The system has been designed for ease of use, even for someone with a minimal un
derstanding of the method, while also leaving the mechanics of the method clear for those
who wish to learn it.
The objective to produce fabrication and design information has also been successfully
integrated into the package, leading the user from the stage of forming and analysing a
dome model through to using the results of the analysis to design safe joint connections.
A linear relationship was found between material properties and estimated failure
strength of the dome’s joints. Though this appeared to correlate well, time limitations
prevented a thorough testing of the inﬂuence of disc radius as a variable.
The physical tension tests performed were good for comparative, empirical studies, and
the results conﬁrmed the theoretical relationship developed was a good match. However,
the accuracy of the values was questionable as they did not reproduce real conditions, and
more work would ideally need to be carried out to conﬁrm the validity of the relationship.
The results of the Strand7 analysis were indicative of the problem in the Water Bank
project using subterranean domes; very large compression forces became prevalent in the
structure, especially in cases of asymmetric loading.
7.1 Further work recommendations
• Expansion of the spreadsheet method to consider more dome conﬁgurations. Cur
rently the analysis focuses on the most common conﬁguration (V3 5/8ths), but with
work this could be expanded, oﬀering more ﬂexibility in terms of fabrication and
48
the size of structure that can be produced
27
.
• In depth investigation of the Water Bank project subterranean loading case. Though
the analysis and designs tools discussed in this dissertation will produce a safe
design for a subterranean dome, conservative assumptions were made to simplify
the loading and the capacity of the structure that may need to be reviewed for this
special case. For example, the contribution of any passive resistance in the soil was
neglected and the testing of the joints focused on tensile rather than compressive
failure.
• Work on improving the eﬃciency of the spreadsheet solver. For higher order prob
lems, it may be necessary to investigate the implementation of node ordering algo
rithms to minimise matrix bandwidth [Cuthill and McKee, 1969].
• Investigation into the design of the dome to improve load bearing capacity. For
example, discs were the only joint considered as they are the design which Vigyan
Ashram currently use and are set up to manufacture. Alternative joint designs may
prove to be cheaper, easier to fabricate or stronger.
• The modiﬁcation of the spreadsheet analysis method to work on a freeware package,
such as OpenOﬃce
28
, so that the software cost associated with the program is
eliminated entirely.
27
As member lengths for lower frequency domes will fail in buckling if particularly large radius domes
are planned; this limits the size of dome that can be achieved using a V3 dome frequency.
28
http://www.openoﬃce.org/
49
References
N.M. Bhandari, P. Krishna, and K. Kumar. IS875: Wind loads on build
ings and structures. [Last Accessed: 28th February 2009], 1987. URL
http://www.iitk.ac.in/nicee/IITKGSDMA/W02.pdf.
N.M. Bhandari, P. Krishna, and K. Kumar. Wind storms, damage and guide
lines for mitigative measures. [Last Accessed: 28th February 2009], 2003. URL
http://www.iitk.ac.in/nicee/IITKGSDMA/W03.pdf.
E. J. Billo. Excel for Scientists and Engineers – Numerical Methods. John Wiley and
Sons, 2007.
BSI. Structural Eurocodes. MFK Group, second edition, 2007.
Robert William Burkhardt. A practical guide to tensegrity de
sign. [Last Accessed: 19th February 2009], 2007. URL
http://bobwb.tripod.com/tenseg/book/revisions.html.
R.C. Coates, M.G Coutie, and F.K. Kong. Structural Analysis. Chapman and Hall, third
edition, 1988.
Fiona Cobb. Structural Engineer’s Pocket Book. Blackwell House, 2007.
R. F. Craig. Craig’s Soil Mechanics. Spon Press, seventh edition, 2004.
E. Cuthill and J. McKee. Reducing the bandwidth of sparse symmetric matrices. In Proc.
24th Nat. Conf. ACM, pages 157–172, 1969.
T. Davis. Gedoesic domes. [Last Accessed: 18th January 2009], 2007. URL
http://www.geometer.org/mathcircles/geodesic.pdf.
D. J. Dowrick. Earthquake Resistant Design  A manual for Engineers and Architects.
John Wiley and Sons, Ltd., 1978.
H. Graham Flegg. Boolean Algebra and its application. Blackie, University of California,
1965.
J. P. Hartog. Understanding ferrocement construction. VITA, Arlington, Virginia, 1984.
Indian Express. No geodesic structures for quake site. Newspaper Article, October 13th
1993.
Infobase. Natural hazard map of india. [Last Accessed: 1st March 2009], 2000. URL
http://www.mapsofindia.com/maps/india/naturalhazard.htm.
50
S. Kalbag. Vigyan Ashram: Rural development education system. [Last Accessed: 18th
January 2009], 2004. URL http://www.vigyanashram.com/.
M. Kardysz, J. Rebielak, and R. Tarczewski. Loading behaviour of some types of tension
strut domes. Space Structures 5, 2:1209–1218, 2002.
H. Kenner. Geodesic math and how to use it. University of California Press, second
edition, 2003.
K. Knebel, J. SanchezAlvarez, and S. Zimmermann. The structural making of the Eden
domes. Space Structures 5, 1:245–254, 2002.
Erwin Kreyszig. Advanced Engineering Mathematics. John Wiley and Sons, Inc., ninth
edition, 2006.
Tara Landry. Dome formulas. [Last Accessed: 8th March 2009], 2002. URL
http://www.desertdomes.com/formula.html.
Z. S. Makowski. Steel space structures. Building with Steel, 11:2–10, 1972.
Z. S. Makowski. Steel Space Structures. Michael Joseph, Battersea College of Technology,
London, UK, 1965.
Z. S. Makowski. Braced domes. Building Speciﬁcation October Edition, pages 37–46, 1979.
Z. S. Makowski. Analysis, Design and Construction of Braced Domes. Granada, University
of Surrey, UK, 1984.
MapXL. Annual temperature map of india. [Last Accessed: 27th February 2009], 2006.
URL http://www.mapxl.com/indiamaps/annualtemperature.html.
Bernard Massey. Mechanics of Fluids. Spon Press, seventh edition, 1998.
J.L. Meek and S. Loganathan. Large displacement analysis of space frame structures.
Computer methods in applied mechanics and engineering, 72:57–75, 1989.
W. Morgan. The Elements of Structure. Pitman, University of Auckland, New Zealand,
second edition, 1981.
R. Motro. Review of the development of geodesic domes (ed. Makowski). Analysis, Design
and Construction of Braced Domes, pages 387–412, 1984.
P. Mullord. Introduction to the analysis of braced domes (ed. Makowski). Analysis,
Design and Construction of Braced Domes, pages 86–95, 1984.
P.D. Pakandam and B. Sarshar. Comparison of the behaviour of 3 types of braced dome.
Space Structures 4, 1:359–368, 1993.
51
C. Pearson. Array formulas. [Last Accessed: 23rd February 2009], 2008. URL
http://www.cpearson.com/excel/ArrayFormulas.aspx.
V. Quintas and J. M. Avila. An analysis of braced domes. Space Structures 4, 1:265–274,
1993.
G. Ranjan and A.S.R Rao. Basic and Applied Soil Mechanics. New Age International
Publishers, second edition, 2000.
D. L. Richter. Space structures development from early concept to temcor geodesic domes.
2nd International Space Structures Conference, pages 534–549, 1975.
Horst Friedrich Rolly. Earthquake Disaster Management. Peter Lang, Frankfurt, Ger
many, 2007.
J.S. SanchezAlvarez. Introduction to the analysis of braced domes (ed. Makowski). Anal
ysis, Design and Construction of Braced Domes, pages 175–244, 1984.
A.V. Shroﬀ and D.L. Shah. Soil Mechanics and Geotechnical Engineering. Taylor and
Francis, 2003.
K. A. Stroud. Engineering Mathematics. Palgrave Macmillan, ﬁfth edition, 2001.
Kian Teh and Laurie Morgan. The application of excel in teaching ﬁnite element analysis
to ﬁnal year engineering students. Proceedings of the 2005 ASEE/A aeE 4th Global
Colloquium on Engineering Education, 2005.
Jon Trevelyan. Level 3 stress analysis. [Course lecture notes], 2007.
Thomas T. K. Zung. Buckminster Ful ler: Anthology for the New Mil lennium. St. Martin’s
Press, New York, USA, 2001.
52
A Project Gantt chart
Figure 31: Project Gantt chart
B Joint forcedisplacement curves
Figure 32: Stress analysis forcedisplacement plot for 1.5mm disc
53
Figure 33: Stress analysis forcedisplacement plot for 2.0mm disc
Figure 34: Stress analysis forcedisplacement plot for 3.0mm disc
54
By.
Basic wind speed (clause 5.2)
From figure 1, the Indian windspeed map gives: V
b
= m/s
Design wind speed (clause 5.3)
Where:
Assuming a standard design life of 50 years,
k
1
= (Table 1, clause 5.3.1)
Assume that Pabal dome is built in category 2 open terrain, with well scattered obstructions:
k
2
= (Table 2, clause 5.3.1)
Assuming no hills, cliffs or enscarpments that channel the wind are nearby:
k
3
= (clause 5.3.3)
k
4
= (clause 5.3.4)
COPYRIGHT ©
Checked.
Wind Loading
1.0
1.0
39.0
1.0
1.0
Structural Analysis of Geodesic Domes
Project.
Wind calculations for Pabal dome MK
Appendix C
Project No. Sheet No.
001 Marek Kubik (Durham Univeristy)
Date.
Feb 2009
The following calculations determine the basic windspeed velocity for a geodesic dome structure built in an
area nearby to Pabal, India. The calculations are guidelines from the Indian Standards, IS875 (Part 3): 1987,
Wind Loads for Buildings and Structures. The relevant clauses to each calculation are given, and any
assumptions are stated throughout.
Basic wind speed is based on peak gust speed averaged over a short time interval of about 3 seconds and
corresponds to 10m height above the mean ground level in an open terrain (Category 2).
Design wind speed modifies the basic wind speed to account for terrain effects and the actual height of the
building under consideration:
4 3 2 1
k k k k V V
b z
Vz = design wind speed at any height z in m/s,
k
1
= probability factor (risk coefficient) (see 5.3.1),
k
2
= terrain roughness and height factor (see 5.3.2),
k
3
= topography factor (see 5.3.3)
k
4
= importance factor for the cyclonic region (see 5.3.4).
NOTE: The wind speed may be taken as constant upto a height of 10 m. However, pressures for
buildings less than 10m high may be reduced by 20% for stability and design of the framing.
Assuming the structure is of normal importance (non‐industrial and not of post cyclonic importance, such
as a cyclone shelter, community building or water tank):
By.
hence, assuming the height of the structure does not exceed 10m:
= x x x x x = m/s
Design wind pressure (clause 5.4)
Where:
hence,
p
z
=
The design wind pressure pd can be obtained as,
For circular or near–circular forms, such as the Pabal dome, the wind directionality factor is:
K
d
= (clause 5.4.1)
K
d
= (clause 5.4.1)
K
c
= (Table 19, clause 6.2.3.14)
hence,
p
d = x x x =
COPYRIGHT ©
V
z
39.0 1.0 1.0 1.0 1.0 0.8 31.2
1.0
0.584
1.0
N/m
2
1.0
Sheet No.
002
1.0 N/m
2
Checked.
MK
Appendix C
Project No.
1.0 1.0 0.584 0.584
Structural Analysis of Geodesic Domes
Project.
Wind calculations for Pabal dome
Marek Kubik (Durham Univeristy)
Date.
Feb 2009
The wind pressure at any height above mean ground level is obtained by the following relationship
between wind pressure and wind speed:
2
6 . 0
z z
V p
p
z
= wind pressure in N/m
2
at height z, and
V
z
= design wind speed in m/s at height z.
Where:
K
d
= Wind directionality factor
K
a
= Area averaging factor
K
c
= Combination factor (see 6.2.3.13)
z c a d d
p k k k p
Pressure coefficients are a result of averaging the measured pressure values over a given area. As the area
becomes larger, the correlation of measured values decrease and vice‐versa. The decrease in pressures due
to larger areas may be accounted for with a reduction factor. Assume that no reduction factor applies to
Pabal dome for simplicty, otherwise design pressure will become dependant on geometry.
Combining factors for wind pressure contributed from two or more building surfaces can allow a reduction
factor because wind pressures fluctuate greatly and do not occur simultaneously on all building surfaces.
However, assuming wind action on the Pabal dome acts in only one direction (worst case):
B
y
.
1
.
I
n
t
r
o
d
u
c
t
i
o
n
2
.
D
o
m
e
G
e
o
m
e
t
r
y
4
3
.
F
o
r
c
e
s
a
p
p
l
i
e
d
S
e
l
f
W
e
i
g
h
t
C
O
P
Y
R
I
G
H
T
©
k
g
/
m
2
E
l
e
m
u
n
i
t
w
e
i
g
h
t
J
o
i
n
t
w
e
i
g
h
t
C
o
a
r
s
e
m
e
s
h
7
7
.
2
1
1
.
1
5
k
N
/
m
3
k
g
p
e
r
j
o
i
n
t
M
K
C
h
e
c
k
e
d
.
2
5
x
2
5
x
3
D
o
m
e
r
a
d
i
u
s
3
m
E
x
p
a
n
d
b
a
s
e
t
o
r
a
d
i
u
s
?
y
e
s
2
2
0
.
0
5
2
.
0
0
E
+
1
1
D
o
m
e
t
y
p
e
V
3
5
/
8
t
h
s
C
l
a
d
d
i
n
g
d
e
n
s
i
t
y
C
l
a
d
d
i
n
g
t
h
i
c
k
n
e
s
s
D
a
t
e
.
V
i
g
y
a
n
A
s
h
r
a
m
P
r
o
j
e
c
t
N
o
.
S
h
e
e
t
N
o
.
0
9

0
0
1
0
0
1
P
a
b
a
l
D
o
m
e
F
a
b
r
i
c
a
t
i
o
n
D
e
t
a
i
l
s
M
a
r
e
k
K
u
b
i
k
(
D
u
r
h
a
m
U
n
i
v
e
r
s
i
t
y
)
N
/
m
2
F
e
b
2
0
0
9
P
r
o
j
e
c
t
.
A
p
p
e
n
d
i
x
E

E
x
a
m
p
l
e
C
a
l
c
u
l
a
t
i
o
n
s
k
N
/
m
3
m
E
l
e
m
s
e
c
t
i
o
n
t
y
p
e
F
i
n
e
m
e
s
h
0
.
9
3
k
g
/
m
2
Y
o
u
n
g
'
s
M
o
d
u
l
u
s
E
T
h
e
f
o
l
l
o
w
i
n
g
d
o
c
u
m
e
n
t
p
r
o
v
i
d
e
s
f
a
b
r
i
c
a
t
i
o
n
d
e
t
a
i
l
s
f
o
r
t
h
e
P
a
b
a
l
D
o
m
e
d
e
s
i
g
n
e
d
b
y
V
i
g
y
a
n
A
s
h
r
a
m
,
b
a
s
e
d
o
n
c
a
l
c
u
l
a
t
i
o
n
s
t
h
a
t
p
r
o
v
i
d
e
n
u
m
e
r
i
c
a
l
p
r
o
o
f
o
f
a
d
e
q
u
a
c
y
o
f
t
h
e
c
o
n
n
e
c
t
i
o
n
d
e
t
a
i
l
s
.
T
h
e
s
t
r
u
c
t
u
r
e
w
a
s
a
n
a
l
y
s
e
d
u
s
i
n
g
a
F
i
n
i
t
e
E
l
e
m
e
n
t
A
n
a
l
y
s
i
s
p
a
c
k
a
g
e
,
t
h
e
m
o
s
t
s
i
g
n
i
f
i
c
a
n
t
p
a
r
a
m
e
t
e
r
s
a
n
d
r
e
s
u
l
t
s
o
f
w
h
i
c
h
a
r
e
s
u
m
m
a
r
i
s
e
d
b
e
l
o
w
.
T
h
e
c
o
n
n
e
c
t
i
o
n
d
e
s
i
g
n
,
w
h
e
r
e
p
o
s
s
i
b
l
e
,
f
o
l
l
o
w
s
E
u
r
o
c
o
d
e
3
,
w
i
t
h
a
l
l
f
o
r
m
u
l
a
e
a
n
d
t
a
b
l
e
s
r
e
f
e
r
e
n
c
e
d
w
h
e
r
e
t
h
i
s
i
s
t
h
e
c
a
s
e
.
C
o
n
t
e
n
t
s
1
.
I
n
t
r
o
d
u
c
t
i
o
n
2
.
D
o
m
e
G
e
o
m
e
t
r
y
3
.
F
o
r
c
e
s
A
p
p
l
i
e
d
4
.
S
u
m
m
a
r
y
o
f
F
E
A
A
n
a
l
y
s
i
s
5
.
C
a
l
c
u
l
a
t
i
o
n
s
6
.
F
a
b
r
i
c
a
t
i
o
n
D
a
t
a
7
.
A
p
p
e
n
d
i
x
T
h
e
f
o
l
l
o
w
i
n
g
f
o
r
c
e
s
c
o
n
t
r
i
b
u
t
e
d
t
o
t
h
e
s
e
l
f
w
e
i
g
h
t
o
f
t
h
e
s
t
r
u
c
t
u
r
e
.
T
h
e
s
e
c
o
m
p
o
n
e
n
t
s
i
n
c
l
u
d
e
a
l
l
t
h
e
t
r
u
s
s
e
l
e
m
e
n
t
s
,
j
o
i
n
t
s
a
n
d
t
h
e
w
e
i
g
h
t
o
f
c
l
a
d
d
i
n
g
:
B
y
.
I
m
p
o
s
e
d
a
n
d
W
i
n
d
L
o
a
d
i
n
g
S
o
i
l
L
o
a
d
i
n
g
L
o
a
d
C
o
m
b
i
n
a
t
i
o
n
F
a
c
t
o
r
s
4
.
S
u
m
m
a
r
y
o
f
F
E
A
a
n
a
l
y
s
i
s
@ @ @
b
e
a
m
@
b
e
a
m
1
3
k
N
C
O
P
Y
R
I
G
H
T
©
W
i
n
d
I
m
p
o
s
e
d
1
8
0
0
.
7
5
1
.
0
5
P
S
F
1
.
3
5
0
.
0
0
L
o
a
d
S
e
l
f
W
e
i
g
h
t
1
6
4
k
N
1
3
1
B
l
a
c
k
C
o
t
t
o
n
D
O
F
m
m
m
m
4
5
1
L
a
r
g
e
s
t
t
e
n
s
i
l
e
f
o
r
c
e
:
L
a
r
g
e
s
t
c
o
m
p
r
e
s
s
i
v
e
f
o
r
c
e
:
D
O
F
d
e
g
L
a
g
e
s
t
p
o
s
i
t
i
v
e
d
i
s
p
l
a
c
e
m
e
n
t
:
I
m
p
o
s
e
d
l
o
a
d
S
o
i
l
0
.
7
8
W
i
n
d
f
o
r
c
e
S
o
i
l
t
y
p
e
S
o
i
l
d
e
p
t
h
A
n
g
l
e
b
u
r
i
e
d
C
h
e
c
k
e
d
.
m
0
.
6
3
.
5
9
k
N
/
m
3
M
K
7
.
2
2
V
i
g
y
a
n
A
s
h
r
a
m
P
r
o
j
e
c
t
N
o
.
S
h
e
e
t
N
o
.
0
9

0
0
1
0
0
2
D
a
t
e
.
F
e
b
2
0
0
9
M
a
r
e
k
K
u
b
i
k
(
D
u
r
h
a
m
U
n
i
v
e
r
s
i
t
y
)
k
N
P
a
b
a
l
D
o
m
e
F
a
b
r
i
c
a
t
i
o
n
D
e
t
a
i
l
s
P
r
o
j
e
c
t
.
A
p
p
e
n
d
i
x
E

E
x
a
m
p
l
e
C
a
l
c
u
l
a
t
i
o
n
s

1
.
5
1
7
.
7

6
.
0
L
a
r
g
e
s
t
n
e
g
a
t
i
v
e
d
i
s
p
l
a
c
e
m
e
n
t
:
W
i
n
d
l
o
a
d
i
n
g
r
e
f
e
r
s
t
o
t
h
e
t
o
t
a
l
f
o
r
c
e
a
c
t
i
n
g
o
n
a
g
i
v
e
n
f
a
c
e
t
h
a
t
w
a
s
c
o
n
s
i
d
e
r
e
d
i
n
t
h
e
d
e
s
i
g
n
,
a
s
d
e
t
e
r
m
i
n
e
d
b
y
t
h
e
w
i
n
d
l
o
a
d
i
n
g
c
o
d
e
I
S
8
7
5
(
P
a
r
t
3
)
.
I
m
p
o
s
e
d
l
o
a
d
i
n
g
i
s
a
n
a
l
l
o
w
a
n
c
e
f
o
r
n
o
n

p
e
r
m
a
n
e
n
t
f
o
r
c
e
s
(
s
u
c
h
a
s
w
o
r
k
e
r
a
c
c
e
s
s
f
o
r
c
l
e
a
n
i
n
g
o
r
p
a
i
n
t
i
n
g
)
.
S
o
i
l
l
o
a
d
i
n
g
i
n
d
i
c
a
t
e
s
w
h
e
t
h
e
r
t
h
e
d
o
m
e
w
a
s
a
n
a
l
y
s
e
d
u
n
d
e
r
c
o
n
d
i
t
i
o
n
s
o
f
p
a
r
t
i
a
l
o
r
f
u
l
l
b
u
r
i
a
l
u
n
d
e
r
s
o
m
e
t
y
p
e
o
f
s
o
i
l
:
L
o
a
d
c
o
m
b
i
n
a
t
i
o
n
f
a
c
t
o
r
s
a
l
l
o
w
t
h
e
u
s
e
r
t
o
w
e
i
g
h
t
t
h
e
a
f
o
r
e
m
e
n
t
i
o
n
e
d
l
o
a
d
s
b
e
f
o
r
e
t
h
e
y
a
r
e
c
o
m
b
i
n
e
d
t
o
f
o
r
m
a
c
o
m
b
i
n
a
t
i
o
n
l
o
a
d
c
a
s
e
t
h
a
t
i
s
u
s
e
d
i
n
t
h
e
F
E
A
a
n
a
l
y
s
i
s
(
f
a
c
t
o
r
s
o
f
z
e
r
o
i
n
d
i
c
a
t
e
t
h
e
t
y
p
e
o
f
l
o
a
d
w
a
s
i
g
n
o
r
e
d
)
.
T
h
e
f
o
l
l
o
w
i
n
g
s
e
c
t
i
o
n
p
r
o
v
i
d
e
s
a
s
u
m
m
a
r
y
o
f
s
o
m
e
o
f
t
h
e
k
e
y
r
e
s
u
l
t
s
d
e
t
e
r
m
i
n
e
d
b
y
t
h
e
F
E
A
s
o
l
v
e
r
:
B
y
.
5
.
C
a
l
c
u
l
a
t
i
o
n
s
k
N
D
i
s
c
j
o
i
n
t
p
a
r
a
m
e
t
e
r
s
C
O
P
Y
R
I
G
H
T
©
S
2
7
5
4
.
6
2
P
r
o
j
e
c
t
.
A
p
p
e
n
d
i
x
E

E
x
a
m
p
l
e
C
a
l
c
u
l
a
t
i
o
n
s
M
K
V
i
g
y
a
n
A
s
h
r
a
m
M
a
r
e
k
K
u
b
i
k
(
D
u
r
h
a
m
U
n
i
v
e
r
s
i
t
y
)
S
h
e
e
t
N
o
.
0
9

0
0
1
0
0
3
D
a
t
e
.
F
e
b
2
0
0
9
P
r
o
j
e
c
t
N
o
.
P
a
b
a
l
D
o
m
e
F
a
b
r
i
c
a
t
i
o
n
D
e
t
a
i
l
s
7
.
7
S
t
e
e
l
t
y
p
e
D
i
s
c
t
h
i
c
k
n
e
s
s
t
m
m
B
o
l
t
c
l
a
s
s
n
o
1
0
2
N
o
.
B
o
l
t
s
p
e
r
c
o
n
n
e
c
t
i
o
n
m
m
C
o
n
t
r
o
l
l
e
d
b
o
l
t
s
i
z
e
C
h
e
c
k
e
d
.
B
o
l
t
d
i
a
m
e
t
e
r
d
0
T
h
e
f
o
l
l
o
w
i
n
g
d
e
s
i
g
n
c
h
e
c
k
s
a
r
e
m
a
d
e
i
n
t
h
e
s
e
c
a
l
c
u
l
a
t
i
o
n
s
:
1
.
C
a
p
a
c
i
t
y
o
f
d
o
m
e
j
o
i
n
t
a
g
a
i
n
s
t
s
h
e
a
r
i
n
g
f
a
i
l
u
r
e
.
2
.
C
a
p
a
c
i
t
y
o
f
b
o
l
t
e
d
c
o
n
n
e
c
t
i
o
n
a
g
a
i
n
s
t
s
h
e
a
r
a
n
d
b
e
a
r
i
n
g
f
a
i
l
u
r
e
.
3
.
C
a
p
a
c
i
t
y
o
f
e
q
u
a
l
a
n
g
l
e
c
h
a
n
n
e
l
s
e
c
t
i
o
n
u
s
e
d
a
s
m
e
m
b
e
r
.
4
.
D
e
f
l
e
c
t
i
o
n
s
a
r
e
n
o
t
e
x
c
e
s
s
i
v
e
u
n
d
e
r
S
e
r
v
i
c
a
b
i
l
i
t
y
L
i
m
i
t
S
t
a
t
e
(
S
L
S
)
.
I
t
i
s
a
s
s
u
m
e
d
t
h
a
t
t
h
e
c
r
i
t
i
c
a
l
c
a
s
e
f
o
r
d
e
s
i
g
n
w
i
l
l
a
l
w
a
y
s
b
e
t
e
n
s
i
l
e
i
n
t
h
e
s
t
r
u
c
t
u
r
e
(
i
.
e
.
a
n
y
f
o
r
c
e
t
h
a
t
c
a
n
b
e
w
i
t
h
s
t
o
o
d
i
n
t
e
n
s
i
o
n
w
i
l
l
a
l
s
o
p
a
s
s
a
n
y
c
o
m
p
r
e
s
s
i
o
n
c
h
e
c
k
s
)
,
s
o
t
h
e
m
o
d
u
l
u
s
o
f
t
h
e
g
r
e
a
t
e
s
t
f
o
r
c
e
a
s
d
e
t
e
r
m
i
n
e
d
b
y
t
h
e
F
E
A
S
o
l
v
e
r
i
s
t
a
k
e
n
a
s
t
h
e
u
l
t
i
m
a
t
e
d
e
s
i
g
n
l
o
a
d
:
D
e
s
i
g
n
A
s
s
u
m
p
t
i
o
n
s
1
.
J
o
i
n
t
s
a
r
e
c
o
n
s
t
r
u
c
t
e
d
f
r
o
m
i
s
o
t
r
o
p
i
c
m
i
l
d
s
t
e
e
l
.
2
.
B
o
l
t
d
i
a
m
e
t
e
r
s
,
b
o
l
t
h
o
l
e
s
a
n
d
s
p
a
c
i
n
g
s
u
s
e
d
a
r
e
h
o
m
o
g
e
n
o
u
s
(
t
h
e
s
a
m
e
)
t
h
r
o
u
g
h
o
u
t
t
h
e
s
t
r
u
c
t
u
r
e
.
3
.
T
h
e
d
i
s
c
d
i
a
m
e
t
e
r
i
s
7
7
.
5
m
m
;
e
x
t
r
e
m
e
c
a
u
t
i
o
n
i
s
r
e
q
u
i
r
e
d
a
l
t
e
r
i
n
g
t
h
i
s
r
a
d
i
u
s
s
i
g
n
i
f
i
c
a
n
t
l
y
,
a
s
t
h
e
j
o
i
n
t
c
a
p
a
c
i
t
y
r
e
l
a
t
i
o
n
s
h
i
p
w
a
s
d
e
s
i
g
n
e
d
w
i
t
h
t
h
i
s
a
s
a
f
i
x
e
d
p
a
r
a
m
e
t
e
r
.
4
.
M
e
m
b
e
r
s
a
r
e
s
u
f
f
i
c
i
e
n
t
l
y
s
q
u
a
t
n
o
t
t
o
f
a
i
l
i
n
b
u
c
k
l
i
n
g
,
a
n
d
b
e
n
d
i
n
g
m
o
m
e
n
t
s
i
n
t
h
e
s
t
r
u
c
t
u
r
e
a
r
e
n
e
g
l
i
g
i
b
l
e
.
5
.
B
o
l
t
s
a
r
e
a
s
s
u
m
e
d
c
a
t
e
g
o
r
y
A
:
B
e
a
r
i
n
g
t
y
p
e
,
i
n
a
c
c
o
r
d
a
n
c
e
t
o
E
C
3
a
n
d
a
r
e
d
e
s
i
g
n
e
d
t
o
w
i
t
h
s
t
a
n
d
p
r
e
d
o
m
i
n
a
n
t
l
y
s
h
e
a
r
i
n
g
a
c
t
i
o
n
(
d
u
e
t
o
a
x
i
a
l
l
o
a
d
i
n
g
)
.
=
E
d
v
F
,
B
y
.
B
o
l
t
s
p
a
c
i
n
g
s
L
i
m
i
t
s
:
J
o
i
n
t
C
a
p
a
c
i
t
y
W
h
e
r
e
:
(
a
s
a
f
e
t
y
f
a
c
t
o
r
o
n
t
h
e
l
o
a
d
)
C
h
e
c
k
t
h
a
t
d
i
s
c
j
o
i
n
t
c
a
p
a
c
i
t
y
s
u
r
p
a
s
s
e
s
t
h
e
u
l
t
i
m
a
t
e
d
e
s
i
g
n
l
o
a
d
:
t
=
1
2
≤
p
1
≥
1
2
≤
p
2
≥
2
f
u
=
γ
=
P
r
o
j
e
c
t
.
S
p
a
c
i
n
g
o
f
b
o
l
t
h
o
l
e
s
f
o
r
m
e
m
b
e
r
s
S
p
a
c
i
n
g
o
f
b
o
l
t
h
o
l
e
s
f
o
r
d
i
s
c
m
m
e
1
e
2
m
m
≤
e
2
1
2
.
5
1
2
.
5
P
r
o
j
e
c
t
N
o
.
M
K
C
h
e
c
k
e
d
.
V
i
g
y
a
n
A
s
h
r
a
m
M
a
r
e
k
K
u
b
i
k
(
D
u
r
h
a
m
U
n
i
v
e
r
s
i
t
y
)
P
a
b
a
l
D
o
m
e
F
a
b
r
i
c
a
t
i
o
n
D
e
t
a
i
l
s
A
p
p
e
n
d
i
x
E

E
x
a
m
p
l
e
C
a
l
c
u
l
a
t
i
o
n
s
7
.
7
0
9

0
0
1
0
0
4
D
a
t
e
.
F
e
b
2
0
0
9
2
2
1
4
0
≤
e
1
2
4
p
1
m
m
N
/
m
m
2
2
5
p
2
4
3
0
1
4
.
2
k
N
>
m
m
1
.
0
5
1
4
0
O
K
C
O
P
Y
R
I
G
H
T
©
2
5
m
m
k
N
S
h
e
e
t
N
o
.
T
h
e
f
o
l
l
o
w
i
n
g
j
o
i
n
t
c
a
p
a
c
i
t
y
r
e
l
a
t
i
o
n
s
h
i
p
w
a
s
d
e
v
e
l
o
p
e
d
e
m
p
i
r
i
c
a
l
l
y
f
o
l
l
o
w
i
n
g
p
h
y
s
i
c
a
l
t
e
n
s
i
o
n
t
e
s
t
s
a
n
d
S
t
r
e
s
s
A
n
a
l
y
s
i
s
o
f
t
h
e
h
u
b
j
o
i
n
t
s
a
t
t
h
e
U
n
i
v
e
r
s
i
t
y
o
f
D
u
r
h
a
m
:
T
h
e
a
r
r
a
y
o
f
b
o
l
t
s
p
a
c
i
n
g
s
m
u
s
t
c
o
n
f
o
r
m
t
o
t
h
e
s
p
a
c
i
n
g
l
i
m
i
t
s
s
h
o
w
n
b
e
l
o
w
o
n
b
o
t
h
t
h
e
d
i
s
c
a
n
d
t
h
e
b
e
a
m
.
I
f
t
h
i
s
i
s
n
o
t
p
o
s
s
i
b
l
e
w
i
t
h
t
h
e
n
u
m
b
e
r
o
f
b
o
l
t
s
c
h
o
s
e
n
,
t
h
e
n
e
i
t
h
e
r
a
h
i
g
h
e
r
g
r
a
d
e
o
f
b
o
l
t
c
l
a
s
s
s
h
o
u
l
d
b
e
c
h
o
s
e
n
o
r
l
a
r
g
e
r
c
o
n
n
e
c
t
i
n
g
m
e
m
b
e
r
s
a
r
e
r
e
q
u
i
r
e
d
.
(
)
γ
4
3
.
2
6
1
3
.
4
,
+
=
t
f
F
u
R
d
v
B
y
.
B
o
l
t
C
a
p
a
c
i
t
y
T
w
o
i
n
e
q
u
a
l
i
t
i
e
s
m
u
s
t
b
e
s
a
t
i
s
f
i
e
d
f
o
r
s
a
f
e
d
e
s
i
g
n
o
f
b
o
l
t
s
:
(
U
l
t
i
m
a
t
e
l
o
a
d
i
s
l
e
s
s
t
h
a
n
d
e
s
i
g
n
s
h
e
a
r
r
e
s
i
s
t
a
n
c
e

E
C
3
c
l
a
u
s
e
3
.
6
)
(
U
l
t
i
m
a
t
e
l
o
a
d
i
s
l
e
s
s
t
h
a
n
d
e
s
i
n
b
e
a
r
i
n
g
r
e
s
i
s
t
a
n
c
e

E
C
3
c
l
a
u
s
e
s
3
.
6
a
n
d
3
.
7
)
S
h
e
a
r
r
e
s
i
s
t
a
n
c
e
c
a
p
a
c
i
t
y
p
e
r
s
h
e
a
r
p
l
a
n
e
:
W
h
e
r
e
:
(
a
s
a
f
e
t
y
f
a
c
t
o
r
f
r
o
m
E
C
3

6
7
S
e
c
t
i
o
n
2
)
B
e
a
r
i
n
g
r
e
s
i
s
t
a
n
c
e
c
a
p
a
c
i
t
y
:
W
h
e
r
e
:
3
x
H
e
n
c
e
:
α
b
=
γ
m
2
=
8
.
0
2
4
3
0
α
b
=
M
I
N
f
u
=
=
1
.
0
0
0
0
.
4
1
7
0
.
9
3
0
1
0
4
0
0
4
3
0
0
.
4
1
7
=
1
.
5
d
=
t
=
1
0
1
2
.
5
m
m
m
m
N
/
m
m
2
P
r
o
j
e
c
t
.
A
p
p
e
n
d
i
x
E

E
x
a
m
p
l
e
C
a
l
c
u
l
a
t
i
o
n
s
(
E
C
3

T
a
b
l
e
3
.
4
)
P
r
o
j
e
c
t
N
o
.
S
h
e
e
t
N
o
.
0
9

0
0
1
0
0
5
D
a
t
e
.
F
e
b
2
0
0
9
V
i
g
y
a
n
A
s
h
r
a
m
M
K
C
h
e
c
k
e
d
.
1
.
5
m
m
2
M
a
r
e
k
K
u
b
i
k
(
D
u
r
h
a
m
U
n
i
v
e
r
s
i
t
y
)
P
a
b
a
l
D
o
m
e
F
a
b
r
i
c
a
t
i
o
n
D
e
t
a
i
l
s
0
.
6
5
0
(
E
C
3

T
a
b
l
e
3
.
4
)
k
N
/
b
o
l
t
4
0
0
N
/
m
m
2
A
=
α
v
=
f
u
b
=
γ
m
2
=
C
O
P
Y
R
I
G
H
T
©
(
A
s
s
u
m
i
n
g
t
h
a
t
α
d
i
s
c
a
l
c
u
l
a
t
e
d
f
o
r
a
n
e
n
d
b
o
l
t
,
w
h
i
c
h
i
s
t
h
e
w
o
r
s
t
c
a
s
e
)
=
R
d
v
,
F
2
1
R
d
b
,
F
m
u
b
d
t
f
k
γ
α
=
2
R
d
v
,
F
m
u
b
v
A
f
γ
α
=
R
d
v
E
d
v
F
F
,
,
≤
R
d
b
E
d
v
F
F
,
,
≤
=
=
0
1
3
d
e
d
α
=
u
u
b
f
f
B
y
.
k
1
=
M
I
N
H
e
n
c
e
:
k
1
=
T
o
d
e
t
e
r
m
i
n
e
t
o
t
a
l
d
e
s
i
g
n
r
e
s
i
s
t
a
n
c
e
o
f
a
c
o
n
n
e
c
t
i
o
n
p
r
o
v
i
d
e
d
b
y
a
n
a
r
r
a
y
o
f
b
o
l
t
s
:
I
f
F
v
,
R
d
≥
F
b
,
R
d
t
h
e
n
:
(
w
h
e
r
e
n
i
s
t
h
e
n
u
m
b
e
r
o
f
b
o
l
t
s
p
e
r
c
o
n
n
e
c
t
i
o
n
)
C
h
e
c
k
t
h
a
t
t
o
t
a
l
d
e
s
i
g
n
r
e
s
i
s
t
a
n
c
e
o
f
b
o
l
t
s
a
r
e
s
u
f
f
i
c
i
e
n
t
:
A
n
g
l
e
S
e
c
t
i
o
n
C
a
p
a
c
i
t
y
A
n
g
l
e
s
e
c
t
i
o
n
:
C
r
o
s
s
S
e
c
t
i
o
n
a
l
A
r
e
a
A
N
e
t
a
r
e
a
a
t
b
o
l
t
h
o
l
e
A
n
e
t
S
e
c
t
i
o
n
t
h
i
c
k
n
e
s
s
t
S
t
e
e
l
g
r
a
d
e
:
1
.
7
k
N
>
k
N
/
b
o
l
t
2
.
8

β
3
4
3
0
k
N
M
K
C
h
e
c
k
e
d
.
O
K
7
.
7
1
.
8
0
0
1
2
.
5
1
0
0
.
4
0
.
5
β
2
f
u
C
O
P
Y
R
I
G
H
T
©
1
5
7
8
.
6
2
.
5
0
0
I
f
F
v
,
R
d
<
F
b
,
R
d
t
h
e
n
:
1
.
8
0
0
4
.
3
P
r
o
j
e
c
t
N
o
.
S
h
e
e
t
N
o
.
0
9

0
0
1
0
0
6
D
a
t
e
.
F
e
b
2
0
0
9
V
i
g
y
a
n
A
s
h
r
a
m
M
a
r
e
k
K
u
b
i
k
(
D
u
r
h
a
m
U
n
i
v
e
r
s
i
t
y
)
P
a
b
a
l
D
o
m
e
F
a
b
r
i
c
a
t
i
o
n
D
e
t
a
i
l
s
P
r
o
j
e
c
t
.
W
a
t
e
r
B
a
n
k
P
r
o
j
e
c
t
C
a
l
c
u
l
a
t
i
o
n
s
=
S
2
7
5
2
5
x
2
5
x
3
m
m
m
m
2
m
m
2
N
/
m
m
2
3
1
2
7
T
h
e
f
i
n
a
l
c
h
e
c
k
U
l
t
i
m
a
t
e
L
i
m
i
t
S
t
a
t
e
c
h
e
c
k
i
s
t
h
a
t
t
h
e
e
q
u
a
l
a
n
g
l
e
s
e
c
t
i
o
n
s
t
h
a
t
m
a
k
e
u
p
t
h
e
t
r
u
s
s
m
e
m
b
e
r
s
w
i
l
l
h
a
v
e
s
u
f
f
i
c
i
e
n
t
r
e
s
i
s
t
a
n
c
e
t
o
t
h
e
a
x
i
a
l
f
o
r
c
e
s
i
n
t
h
e
s
t
r
u
c
t
u
r
e
.
T
h
e
w
e
a
k
e
s
t
p
o
i
n
t
o
f
t
h
e
s
e
m
e
m
b
e
r
s
i
s
t
h
r
o
u
g
h
t
h
e
d
r
i
l
l
e
d
s
e
c
t
i
o
n
,
a
n
d
s
e
v
e
r
a
l
e
q
u
a
t
i
o
n
s
e
x
i
s
t
f
o
r
d
e
t
e
r
m
i
n
i
n
g
t
h
e
s
e
c
t
i
o
n
c
a
p
a
c
i
t
y
,
d
e
p
e
n
d
i
n
g
o
n
t
h
e
n
u
m
b
e
r
o
f
h
o
l
e
s
.
C
h
o
s
e
n
s
e
c
t
i
o
n
s
i
z
e
a
n
d
o
t
h
e
r
r
e
q
u
i
r
e
d
c
a
l
c
u
l
a
t
i
o
n
p
a
r
a
m
e
t
e
r
s
:
=
−
7
.
1
8
.
2
0 2
d e
=
R
d
b
,
F
(
)
n
F
F
R
d
v
R
d
×
=
,
m
i
n
n
F
F
R
d
b
R
d
×
=
,
=
R
d
F
B
y
.
2
.
0
x
(
)
x
x
x x x
C
h
e
c
k
t
h
e
a
n
g
l
e
c
a
p
a
c
i
t
y
i
s
s
u
f
f
i
c
i
e
n
t
:
D
e
f
l
e
c
t
i
o
n
L
i
m
i
t
C
h
e
c
k
6
.
F
a
b
r
i
c
a
t
i
o
n
d
a
t
a
4
w
a
y
c
o
n
n
e
c
t
i
o
n
j
o
i
n
t
s
5
w
a
y
c
o
n
n
e
c
t
i
o
n
j
o
i
n
t
s
6
w
a
y
c
o
n
n
e
c
t
i
o
n
j
o
i
n
t
s
O
K
(
E
C
3
7
.
2
.
1
)
A
r
e
a
l
l
m
e
m
b
e
r
d
e
f
l
e
c
t
i
o
n
s
w
i
t
h
i
n
t
h
e
l
i
m
i
t
(
L
/
3
6
0
)
?
A
s
s
e
m
b
l
y
d
i
a
g
r
a
m
o
f
t
r
u
s
s
e
l
e
m
e
n
t
s
H
I
0
.
0
0
0
0
.
0
0
0
0
.
0
0
0
0
.
0
0
0
k
N
=
4
3
0
1
2
.
9
1
2
7
x
(
2
b
o
l
t
s
)
(
3
+
b
o
l
t
s
)
4
3
0
k
N
=
1
2
.
5
1
.
5
1
0
0
0
0
.
4
4
3
0
1
2
7
x
x
1
0
0
0
C
O
P
Y
R
I
G
H
T
©
=
k
N
1
8
.
2
1
.
5
1
0
0
0
=
0
.
5
x

0
.
5
x
1
0
(
1
b
o
l
t
)
1
.
5
3
0
9

0
0
1
S
h
e
e
t
N
o
.
P
a
b
a
l
D
o
m
e
F
a
b
r
i
c
a
t
i
o
n
D
e
t
a
i
l
s
P
r
o
j
e
c
t
.
A
p
p
e
n
d
i
x
E

E
x
a
m
p
l
e
C
a
l
c
u
l
a
t
i
o
n
s
M
K
V
i
g
y
a
n
A
s
h
r
a
m
M
a
r
e
k
K
u
b
i
k
(
D
u
r
h
a
m
U
n
i
v
e
r
s
i
t
y
)
D
a
t
e
.
F
e
b
2
0
0
9
P
r
o
j
e
c
t
N
o
.
C
h
e
c
k
e
d
.
k
N
O
K
1
4
.
6
1
4
.
6
k
N
>
7
.
7
G
L
e
n
g
t
h
(
m
)
1
.
0
6
3
1
.
2
3
0
1
.
2
5
7
0
0
7
0
.
0
0
0
0
.
0
0
0
M
e
m
b
e
r
A B C D E F
0 0 0 6
1
5
0 0
N
o
.
R
e
q
u
i
r
e
d
3
0
5
5
8
0
0 = =
4
0
T
h
e
d
e
f
l
e
c
t
i
o
n
l
i
m
i
t
c
h
e
c
k
i
s
a
f
i
n
a
l
c
h
e
c
k
t
o
e
n
s
u
r
e
t
h
e
s
i
z
e
o
f
d
e
f
l
e
c
t
i
o
n
s
i
n
t
h
e
s
t
r
u
c
t
u
r
e
u
n
d
e
r
t
h
e
l
o
a
d
i
n
g
c
o
n
d
i
t
i
o
n
s
a
s
s
u
m
e
d
w
i
l
l
n
o
t
b
e
e
x
c
e
s
s
i
v
e
e
n
o
u
g
h
t
o
l
e
a
d
t
o
c
r
a
c
k
i
n
g
o
r
v
i
s
u
a
l
d
i
s
t
o
r
t
i
o
n
s
.
I
f
t
h
i
s
c
h
e
c
k
f
a
i
l
s
,
b
u
t
t
h
e
t
h
r
e
e
U
l
t
i
m
a
t
e
L
i
m
i
t
S
t
a
t
e
c
h
e
c
k
s
h
a
v
e
b
e
e
n
p
a
s
s
e
d
,
t
h
e
n
t
h
e
s
t
r
u
c
t
u
r
e
i
s
s
t
i
l
l
s
t
r
u
c
t
u
r
a
l
l
y
s
o
u
n
d
.
=
R
d
F
(
)
2
0
2
,
5
.
0
0
.
2
m
u
R
d
u
t
f
d
e
N
γ −
=
2
2
,
m
u
n
e
t
R
d
u
f
A
N
γ
β
=
2
3
,
m
u
n
e
t
R
d
u
f
A
N
γ
β
=
B
y
.
7
.
A
p
p
e
n
d
i
x
G
e
o
d
e
s
i
c
d
o
m
e
c
h
o
r
d
f
a
c
t
o
r
s
(
u
n
s
c
a
l
e
d
)
G
e
o
d
e
s
i
c
d
o
m
e
f
a
b
r
i
c
a
t
i
o
n
d
a
t
a
T
a
b
l
e
3
.
1
f
r
o
m
E
u
r
o
c
o
d
e
3

b
o
l
t
c
l
a
s
s
n
o
m
i
n
a
l
s
t
r
e
n
g
t
h
s
E
q
u
a
l
a
n
g
l
e
s
e
c
t
i
o
n
d
a
t
a
E
C
3

3
1
n
o
m
i
n
a
l
s
t
r
e
n
g
t
h
v
a
l
u
e
s
f
o
r
h
o
t
r
o
l
l
e
d
s
t
r
u
c
t
u
r
a
l
s
t
e
e
l
F
G
0
.
5
4
6
5
3
0
.
4
0
3
5
5
V
4
S
h
e
e
t
N
o
. 0
0
8
P
r
o
j
e
c
t
.
A
p
p
e
n
d
i
x
E

E
x
a
m
p
l
e
C
a
l
c
u
l
a
t
i
o
n
s
M
K
C
h
e
c
k
e
d
.
P
a
b
a
l
D
o
m
e
F
a
b
r
i
c
a
t
i
o
n
D
e
t
a
i
l
s
V
5
3
/
8
t
h
s
V
5
5
/
8
t
h
s
V
2
V
3
3
/
8
t
h
s
V
3
5
/
8
t
h
s
1
0
.
9
V
6
M
e
m
b
e
r
L
e
n
g
t
h
D
o
m
e
t
y
p
e
V
1
V
2
V
3
3
/
8
t
h
s
3
5
V
3
5
/
8
t
h
s
2
4
7
V
i
g
y
a
n
A
s
h
r
a
m
P
r
o
j
e
c
t
N
o
.
0
9

0
0
1
M
a
r
e
k
K
u
b
i
k
(
D
u
r
h
a
m
U
n
i
v
e
r
s
i
t
y
)
D
a
t
e
.
F
e
b
2
0
0
9
H
I
B
C
E
6 4
4
0
x
4
0
x
5
3
9
1
5
4
0
x
4
0
x
6
C
O
P
Y
R
I
G
H
T
©
S
3
5
5
S
4
5
0
3
5
5
4
5
0
3
4
0
x
4
0
x
4
V
6
V
5
3
/
8
t
h
s
V
5
5
/
8
t
h
s
D
o
m
e
t
y
p
e
V
4
V
1
0
.
1
6
2
5
6
7
0
.
1
9
0
4
7
7
A
1
.
0
5
1
4
6
0
.
1
9
8
1
4
7
0
.
1
9
8
1
4
7
0
.
6
1
8
0
3
0
.
3
4
8
6
2
0
.
3
4
8
6
2
0
.
2
5
3
1
8
0
.
4
0
3
5
5
0
.
2
9
5
2
4
0
.
2
3
1
7
9
0
.
2
3
1
7
9
0
.
4
1
2
4
1
0
.
4
1
2
4
1
0
.
2
9
4
5
3
0
.
2
2
5
6
8
6
0
.
2
2
5
6
8
6
0
.
1
8
1
9
0
8
D
0
.
3
1
2
8
7
0
.
2
4
7
2
4
3
0
.
2
4
7
2
4
3
0
.
2
0
2
8
2
0
.
3
2
4
9
2
0
.
2
5
5
1
6
7
0
.
2
5
5
1
6
7
0
.
1
8
7
3
8
3
0
.
2
9
8
5
9
0
.
2
4
5
0
8
6
0
.
2
4
5
0
8
6
0
.
1
9
8
0
1
3
0
.
2
6
1
5
9
8
0
.
2
0
5
9
0
8
0
.
2
6
1
5
9
8
0
.
2
3
1
5
9
8
0
.
2
3
1
5
9
8
0
.
2
1
5
3
5
4
0
.
2
4
5
3
4
6
0
.
2
4
5
3
4
6
0
.
2
1
6
6
2
8
0
N
o
.
5
J
5
N
o
.
6
J
2
9
1
2
5
1
5
6
4
0
4
R
e
q
u
i
r
e
d
n
u
m
b
e
r
s
9
0
0
5
1
0
5
5
0
3
2
0
4
6
0
f
u
(
N
/
m
m
2
)
1
0
0
0
2
4
0
3
0
x
3
0
x
4
4
0
x
4
0
x
3
S
2
7
5
2
7
5
4
3
0
3
0
x
3
0
x
5
3
6
0
S
t
e
e
l
G
r
a
d
e
S
2
3
5
f
y
(
N
/
m
m
2
)
2
3
5
3
0
x
3
0
x
3
1
8
7
3 4 5 3
2
4
1
2
5
x
2
5
x
4
6
0
6
0
T
h
i
c
k
n
e
s
s
(
m
m
)
1
3
0
1
2
0
6
0
2
5
x
2
5
x
5
2
5
x
2
5
x
3
2
5
7
0
6
0
f
u
b
(
N
/
m
m
2
)
3
0
2
5
A
n
g
l
e
c
o
d
e
1
5
7
8
0
0
6
.
8
4
0
0
5
0
0
8
.
8
6
0
0
3
0
B
o
l
t
c
l
a
s
s
4
8
0
6
4
0
5
.
6
f
y
b
(
N
/
m
m
2
)
2
4
0
3
0
0
3
0
A
r
e
a
(
m
m
2
)
6
0
3
0
5
0
3
0
6
6
5
1
0
9
0
3
0
4
0
5
5
2
0
0
6
3
0
3
0
8
0
3
0
3
0
3
0
3
0
3
0
N
o
.
A
2
5
3
0
6
0
1
3
0
N
o
.
B
5
0
6
0
3
0
N
o
.
C
3
0
N
o
.
E
N
o
.
D
5
0
6
0
5
0
3
0
N
o
.
F
5
0
6
N
o
.
G
5
N
o
.
H
N
o
.
4
J
N
o
.
I
1
0
6
5
1
0
9
5
6
1
5
6 6
4
.
6
2
0
6
2
0
3
0
3
0
T
h
i
s
a
p
p
e
n
d
i
x
t
a
b
u
l
a
t
e
s
t
h
e
v
a
l
u
e
s
t
h
a
t
h
a
v
e
b
e
e
n
l
o
o
k
e
d
u
p
f
o
r
c
a
l
c
u
l
a
t
i
o
n
s
i
n
v
a
r
i
o
u
s
p
a
r
t
s
o
f
t
h
e
s
p
r
e
a
d
s
h
e
e
t
.
I
t
i
s
n
o
t
n
e
c
e
s
s
a
r
y
t
o
p
r
i
n
t
t
h
i
s
p
a
g
e
.
Abstract A project to house 40 families in the Maharashtra region of India which began in January 2005 was halted shortly after it commenced due to concerns over the loading applied to a series of geodesic domes which form a large portion of the whole complex. Vigyan Ashram, the organisation that manufactures the domes, issued a request for the development of an aﬀordable computer program that would allow them to model the structural response of the domes. The ensuing research into the design of the geodesic domes and the development of a spreadsheet based ﬁnite element package are the subject of this dissertation.
Acknowledgements I wish to thank all at Vigyan Ashram, particularly Yogesh Kulkarni and Ashok Mathur, for their assistance with gathering information relevant to this project. I would also like to express my gratitude to Colin Wintrip and Steven Richardson, whose assistance with fabricating and testing the joints was greatly appreciated. Finally, I wish to thank Dr. Charles Augarde, my supervisor, who has provided me with invaluable guidance and advice throughout this research project.
i
Project plan
Introduction
A project to house 40 families in the Maharashtra region of India which began in January 2005 was halted shortly after it commenced due to concerns over the loading applied to a series of geodesic domes which form a large portion of the whole complex. The domes had soil packed over them, a condition which was never anticipated in the original design. The need for a method to model the structural response of the geodesic or “Pabal” dome was highlighted by Vigyan Ashram, the Non Governmental Organisation (NGO) that manufactures the domes; this dissertation describes the research into the Pabal dome and the development of a bespoke ﬁnite element analysis (FEA) package capable of running in Microsoft Excel.
Objectives
Vigyan Ashram locally produces the geodesic domes as doityourself kits for the lowermiddle class of both rural and urban populations. The original design was adopted in the aftermath of the 1993 Killari earthquake, aiming to provide durable, low cost housing capable of withstanding the earthquakes, rains and winds of India for those that lost their homes. Of the 120 geodesic dome kits supplied to the Water Bank housing project, 40 were planned to be subterranean and were therefore aﬀected. Vigyan Ashram, together with another NGO, engINdia, asked for research into the geodesic dome’s current incarnation to be carried out. The Water Bank project highlighted the need for a method of modelling the structural response of geodesic domes, as an assessment of the loading encountered by the domes would allow recommendations to be made as to how the design could be adapted to accommodate the expected loading. Vigyan Ashram’s main desire was for a structural analysis package without the associated licensing costs, to allow their science and technology centre to assess the geodesic dome’s structural response inhouse. This would provide signiﬁcant long term beneﬁts for Vigyan Ashram, removing the need for outsider aid with such design problems in the future. If appropriate, students at Vigyan Ashram would also use the developed method to further their understanding of structural behaviour. It was hoped that a reliable method of analysis would increase the number of potential uses of the structure and hence its market demand, beneﬁting both the community and the business that manufactures the dome, which was founded and run by an exstudent of Vigyan Ashram. The objectives of this project may be summarised as follows. 1. The development of a method of ﬁnite element analysis for geodesic dome structures ii
overlapping with Phase I. experimental work would be carried out and the spreadsheets developed. how long they are. 2.using Microsoft Excel 2003. In this phase. • Phase III: Project conclusions. Upon return of the draft report from the project supervisor. a relationship between material parameters and the ultimate failure strength of the connections was to be developed. 3. modiﬁcations would be made and the ﬁnal project submitted on the 29th of April. A Gantt chart breakdown of the project plan may be found in Appendix A. including the submission of a risk assessment and the production of engineering drawings to fabricate the required dome joints. Here. Additionally. The program needed to provide a user friendly format that gave the user the opportunity to deﬁne a geometry. preparation work for the later tension testing work would be scheduled. from April until the concluding oral exam in May. culminating in the submission of the literature review. culminating in the submission of a draft version of the ﬁnal report. a poster design for display in the department would be produced and preparation for the oral exams would occur in this period. from December to the end of April. It was required to output information about nodal displacements and forces in the elements. with all the major project deadlines highlighted. The development of an additional spreadsheet that would allow the user to calculate all the relevant dome fabrication details . lasting from September through to January.5 weeks . Project timeline This project was to be carried out over a period of 28. Additionally. The assessment of the dome joints used in the structure to determine their design limits. Work was broadly categorised into three phases. background reading and discussion with Vigyan Ashram would lead to the development of a list of objectives and a project plan. iii . how many bolts are required for suﬃcient connection strength etc. • Phase II: Data collection. Based on these ﬁndings.from the end of September through to midMay. select appropriate material properties and apply a variety of loading conditions. • Phase I: Initial research and planning.how many diﬀerent struts are required.
. . . . . . . . . . . . . . . . .3. . . . . . . . . . .2 Analysis . . . . . . . . . . . . . . . . . . . . . . . . . 2. .3. . . . . . . . . . . . . . . . . 2. . . . . . . . . .3. .1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3 Geodesic domes . . 2. . .4 Lateral earth pressure iv vi vii 1 1 1 2 2 3 4 5 5 5 6 8 8 9 10 10 11 12 12 12 16 17 18 . . . . . . . . . . .2. . . 2. . . . . . . . . . . . . . . . . . . . . .4. . . . . . . . . . . 3. . . . . . . . . . 2. . . . . . . . . . . . . . iii Project timeline . . . . . . . . . . . . . . . . . . . . . . . .3 Post processing . . . . . . . . . . 2. . . . . . . . . . . . . . . . . . . . . 2. . . . . . . . . . . . . . 2. . . . . . . . . . . . . . . . . . . . .4. . . . . . . . . . . . . . .4 Structural analysis . . . 2. . . . . . . . . . 3. . . . . . .2 Vigyan Ashram and engINdia 2. . . . . . . . . . . . . . . . . . .4 Braced dome analysis .1 Buckminster Fuller . . . . . . . 2. . . . . . . iii Objectives . . . . . . . . . . . . . . . . . iv . . . . . . . . . . . . . . . . . . . . . . . . . . .3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii Table of Contents List of Tables List of Figures 1 Introduction 2 Literature review 2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2. . . . . . . . .Contents Abstract Acknowledgements i i Project Plan iii Introduction . . 3 Theory 3. . . .3 The Water Bank project . . . . . . . . . . . . . . . . . . . .3 Explanation of behaviour . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3. . . . . . . . . . . . . . .2. . .2 Pabal . . . . . . . . . . . . . . . . . . . . . . . . . .4 Strengths and weaknesses . . . .5 The Pabal dome . . . . . . . . . . . . .2 Solution methods . . . . . . . .1 Stiﬀness method . . . . . . . . .3. . . . . . . . . . . . . . . . .4. . . . . . . . . . . . . 2. . . . . .2. . . . . .1 Earthquake and aftermath . . . . 2. . . . . . . . . . . . . 2. . . . . . .1 Idealisations . . . . . . 2. . .2 Deﬁnition . . . . . . . . . . . .3 Previous work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5. 5. . . .5. . . . .7 Discussion . . . . . . . . . .4 Load Case V: Imposed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5. . .5. 5. . . . .1 Package structure . . . . . . . . . 4. . . . . . . . 5. . . . 4.5 Combination load cases . . . . . . . . . . . . . . . . . .4 Assumptions . . . . . . . . . . . . . . . . . . . . .2 FEA_Input . . . . . .2 Apparatus . . . . . .1 Introduction to Strand7 .3 Loading . . . . . . . . . . . . . . . . . . . . . . . . . 4. . . . 5.4. .4. . . . . . . . . .6 Discussion . . . . . . . . . . . .3 Soil . . . . . . . . . . . . . . . . . . . . . . . . . . .5. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5 Self weight and imposed loads 4. . . . . . . . . . . . .3 Load Case IV: Wind . . . . . . . . . . . . . . . . . . . . . .4. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5. . . .1 Stiﬀness matrix assembly . . . . . . . . . . . . . . . . . . . . . . 4. . . . . . . . . .5 Results . . . . . . . . . . . . . . . . . . . 4. .3. . . . . . . . . . . . .2 Snow . . . . . . . . .2 Load Cases II & III: Soil . . . . 5. . . . . . . . . . . . . . . . . . . . . .3.3 Discussion . . . . .4. . . . . .5. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4. . . . . . . . . . . . . . 4. .5 Analysis . . . . . . . . . . . . . . . . . . . . . . 5. . . . . . . . . . . . . . . . . . . . . .4. . 6 Spreadsheet development 6. . . . . . 4. . . .1 Introduction . . . . . . .3. . . . . 5. . . 4. . .3 FEA_Solver . . . . . . . . . 5. . . . . . . . . . . . . . . . . . 6. . . . . . . . . . . . . . . .3 Dimensional analysis . . . . . . . . . . . . . .4 Wind . . 5 Joint testing 5. . . .3 Risk assessment . . . . . . . . . . . 4. . . . . . . . . . . . . .4 Tension testing . . . . .1 Load Case I: Self weight . . . . . . . . . . . . . . . . . . . . . . . . . .2 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . 4. . . 4. . . . .1 Joint fabrication . . . . . . . 5. . 5. . .2 Pabal disc joints . .3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4. . . . . . . . . . . .5. . 6. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5. . . . . .4 Geodesic dome analysis 4. . . . . . 4. . . . . . . . . . . . . . . . . . . .2 Coordinate generation . . . . . .5. . . . . . . . . .1 Seismic and impact loads . . . . . .4 Method statement 5. . . . . . . . . . . . . . . . . . . . . . . . . . 6. . . . . . . .6 Results . . . . . . . . . . . . . 18 18 19 20 20 20 21 22 23 25 26 26 26 27 28 28 28 29 29 29 30 31 33 33 33 34 34 35 36 36 36 38 39 40 40 41 41 41 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v . .1 Model development 5. . . . . . . . . . . . . . . .4. . . . . . . . . . . . .3. . . . . . . . .5 Stress analysis . . . . .3. . . . . . . . . . . . . . . . . . . . . . . . . . . . 4. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . Discussion . . . . . . . . . . . . . . . . . . . . . Black Cotton soil properties . . . . . . . . . . . . . . . . . . .5 6. . . . . . 27 28 29 32 32 32 37 39 39 47 vi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Dimensional analysis: step 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 45 45 45 46 7 Conclusions 48 7. . . . . . . . . . . . Convergence test results . . . . . FEA_Output . . . . . . . . . . . . . . . . . . . . . . . . .3. . . . . . . . Dimensional analysis: step 2 . . . . . . . . . . . . . . . . . . . . . 48 References A Project Gantt chart B Joint forcedisplacement curves C Wind loading calculations D Joint engineering drawing E Sample Excel calculation F Risk Assessment and COSSH form 50 53 53 55 56 57 61 List of Tables 1 2 3 4 5 6 7 8 9 10 Element and joint properties . . . . . . . . . . . . . . . . Dimensional analysis: step 1 . . . . . . . . . . . . . . . . .2 Calculations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . capacities . . . . . . . . . . . .4 6. . . . .1 Postprocessing 6. . . . . . . . . . . . . . .4. .4. . 6. . . . . . . . . . . . . . . . . . . . . . . . . . Nonlinear joint analysis results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Comparison of theoretical and practical Spreadsheet based FEA results . . . joint . . . . . . . . . . . . . . . . . .2 Solver . . . . . . . .1 Further work recommendations . . . . .6. . . . . . . . . . . . . . . . . . . . . Strand analysis results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . Sketches of material stressstrain relationships . . . . . . . . . . . . . . . . . Flow chart of design process . . . . . . . . . . . . . . 1984] . . . . . . . . . . . . Stress analysis forcedisplacement plot for 3. . . . Mesh convergence . . . . . . . . . . Assembly of the structure stiﬀness matrix from element stiﬀness matrices. . . . . . . . . . . . Structure stiﬀness matrix bandwidth . . . . . . . . . . . . . . Stress analysis forcedisplacement plot for 1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Structure stiﬀness matrix assembly using SUMPRODUCT . . . . . . . . . . . . . . . . . . . . . Class I subdivision [Motro. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Calculation of member axial forces . . . . . . . . . . . . . .0mm disc . . . . . . .List of Figures 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 Isosahedron . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Stress analysis forcedisplacement plot for 2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Brown University’s 2D FE program . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Assembly of the global stiﬀness matrix . . Regional soil deposits of India (after Shroﬀ and Shah [2003]) . . . . . . Tension test results . Test rigs for joints . . . . 6 6 7 7 7 10 11 14 15 16 19 21 22 23 24 27 30 31 33 34 35 37 37 38 40 42 42 43 44 45 53 53 54 54 vii . . . . . . . . . . . . . . . . . . . Horizontal and vertical stresses acting on a subterranean dome (after Landry [2002]) . . . . . . . . . Project Gantt chart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Example area division for a node . . . . . . . . . Soil loads considered . . . . . . . . . . . . . . . . . Class II subdivision [Motro. . . . . . DOF for a pin jointed element in a local coordinate system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1984] . . . . Close up of Pabal dome joint . . . . . . . Calculation of project member lengths . . . . . . . . . . . . . . . . [1987]) . . . . . . . . . . . . . . . . . . . . . Intermediate matrix assembly step . . . . . . . . . . . Bhandari et al. . . . . . . 1984] . Uniaxial loading of bolt array . . . . . . External wind pressure coeﬃcients for a spherical structure (from Table 18. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1981] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5mm disc . Example subdivision frequencies [Motro. . . . . . . . . . . .0mm disc . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Denison loading machine . . . . . . . . . . . . . . . . . . . Pabal dome . . . . . . Plot of relationship between dimensionless variables . . . . . . . . Great circles [Morgan. . . . . . . . . . . . . . . . . Global coordinate system used for analysis . . . . . . . . . . . .
In the ﬁnal part of the report. Vigyan Ashram. which provided an early basis for gathering results and a benchmark against which to compare the Excel based method.1 Literature review Introduction The purpose of this literature review is to give context to the relevance and necessity of the work undertaken in this dissertation.1 Introduction The project plan included in the prelude of this report deﬁnes the project objectives agreed with Vigyan Ashram and the expected timeline in which they were to be achieved. 2. is then introduced. Previous work in this area is reviewed. Finally. using a combination of physical tension testing and stress analysis. The second part of the literature review examines the deﬁnition of a geodesic dome and looks brieﬂy at its history. An overview of the ﬁnite element (FE) method developed in Excel. 1. with recommendations for further work in this area. a brief history of the geodesic dome’s role in the reconstruction and rehabilitation in wake of the 1993 Killari earthquake is given. and to introduce the concepts and deﬁnitions that will aid the understanding of the rest of the work herein. 3. Firstly. The NGO that has ultimately continued to locally produce the dome kits to the present day. the possible methods of braced dome analysis are reviewed and the decision to produce an Excel based program is justiﬁed. a commercial FEA package. 1 . The main body of the report covers three main aspects of the project. 2 2. The literature review provides context and background information to justify the need for this work and the theory section summarises the necessary information required to understand the FEA method developed in Excel. The strengths and weaknesses of geodesic domes are discussed. The work halted at the Water Bank project in 2005 is used to demonstrate the need for an assessment of the loading on the geodesic domes. The development of a relationship to predict the failure strength of the joints from the material properties and geometry. and the modiﬁed version of the dome used by Vigyan Ashram is introduced. The development of a geodesic dome model and the subsequent analysis of the dome using Strand7. the success of meeting the objectives outlined in the project plan is discussed and conclusions are drawn.
2. These houses performed extremely well. The magnitude of this disaster became apparent only days later.800 livestock were killed. 2007]. spreading word of the tragedy. Local builders reported that “. but the eﬀects of the disaster were felt in more than 2500 villages in the 11 neighbouring districts. 7. .2. who lived in the region at the time of the quake. The press rushed to the scene. The damage was total in 52 villages of the Latur and Osmanabad districts. Ablebodied survivors assisted the injured and dug in debris to rescue the living and recover the dead with their bare hands as initially no excavation equipment was available. often boulders were piled upon shallow foundations. a further 16. ultimately leading to fund raising initiatives.928 human lives lost.and the shortcomings are used to justify Vigyan Ashram’s need for a bespoke structural analysis package. In normal conditions these constructions are relatively stable. Regardless. both national and international. traditionally staying in rooms attached to houses. The Indian armed forces were brought in to aid with the rescue and relief operation. As more and more donations poured into government collection and distribution centres the authorities 2 .1 Pabal Earthquake and aftermath The following description of the earthquake and aftermath are based upon extracts from the account of Professor Horst Rolly [Rolly.” No oﬃcial engineering standards existed in rural Maharashtra before the quake.2 2. but on the soft Black Cotton soil of the region they collapsed easily when the earth shook. The Maharashtra state government in Bombay responded to the news immediately by sending helpers from civic bodies and doctors from neighbouring districts along with supplies. In the early hours of the morning of the 30th September 1993. the Killari earthquake struck the Maharashtra province of India. ﬁnished with mudplaster. with some 10. met the same fate. The structures that suﬀered most were those constructed of thick masonry to provide better insulation against the summer heat.000 were injured and more than 15. the civic response was spontaneous.000 troops bringing in lifting machinery. the failure of stone masonry during the earthquake was largely because of the excessive wall thickness demanded by the Maratha households for thermal comfort and storage of valuables within the thickness of the walls. Animals. . bonded only with minimal cementing. tents and water puriﬁcation units. Real poverty proved to be a blessing as the landless poor lived in thatchedroof buildings made of straw and reeds or split bamboo. No early warning system for such a natural disaster was in place and no comprehensive disaster management plan existed before the quake. The ﬁrst quake lasted for 40 to 50 seconds. survivors recounting that houses “swayed like a cradle” before they caved in and buried people underneath. suﬀering only minor cracks to the mud plaster walls.
so much. Dr Kalbag. Nonconventional structures like geodesic domes were ruled out because people are not used to living in them. bilateral donations from the United Nations Development Program. Vigyan Ashram has continued to manufacture and supply dome kits since the 1993 Killari earthquake. the Indian media reported that politicians were attempting to exclude certain designs: “. Adventist Development and Relief Agency (ADRA) India had already contacted the villagers of Gubal about the possibility of building geodesic domes as a safe housing device. The villagers agreed that a sample geodesic dome could be built at the resettlement site for the villagers to inspect and scrutinise. the design would be adopted as a rehabilitation measure. .” For the longer term reconstruction eﬀort.2 Vigyan Ashram and engINdia Early discussions of ADRA India with the people of Gubal concluded that safe housing was of paramount importance in the villagers’ opinion. . However. the Government of Maharashtra invited NGOs for round table discussions of how to proceed with reconstruction eﬀorts.” [Indian Express. . and none of the surviving inhabitants wished to stay and live in old Gubal. By the time of the round table discussions. Vĳay Kumar and other likeminded engineers based there developed a modiﬁed version of the dome called the “Pabal” dome. . Vigyan Ashram serves as a science centre for rural 3 . no additional help of unsolicited kind is needed either in the form of items or volunteers. As well as a research institution. rather than the 182 originally planned. focusing on making lowcost. .2. until ultimately the Chief Minister publicly announced three days after the quake that “. After careful comparison of earthquake resistant architectural designs. On the condition that the dome met village approval. ADRA India decided upon the geodesic dome. a research institution based in the village of Pabal had experimented with the geodesic dome for a number of years. before the technological aspects of proposals could be fairly assessed. 2. the Asian Development Bank and a World Bank loan granted the Indian Government US$358 million as a low interest emergency loan with a 30 year payback period to add to its resources. October 13th 1993]. the chief minister and his cabinet colleagues are also believed to have taken the psychological aspect into account. The state government wanted to ensure that despite the large variety of agents involved in the reconstruction eﬀort. it is learnt. The Gubal village elder ultimately submitted a request for 365 houses to be built. the houses met common standards of earthquake safety. The horrifying experiences of the Killari earthquake had turned much of the village into a graveyard. which was subsequently marketed by a number of entrepreneurial small scale industries. producing prefabricated components as part of low cost doityourself kits and taking orders to build and maintain the domes. In the third week of October 1993. We already have much. . safe and comfortable housing that is aﬀordable for the lowermiddle class of both urban and rural populations. Vigyan Ashram.became unable to handle them.
Of the 120 domes supplied to the Water Bank housing project 40 are planned to be subterranean and are therefore aﬀected. and home & health.000 at Jan 2009 rates) to house 40 families in Ankoli. measure and record data.net/ http://fab. The course is novel in two aspects: what it achieves through student input and how it beneﬁts the village economy. a signcutter. and also to school dropouts frustrated by conventional education. with the intention of providing Vigyan Ashram and the local community the necessary tools to empower them to solve their own engineering problems.mit. The intention of the project is to promote sustainable housing and fulﬁl the desires of the residents to be selfemployed. energy.3 The Water Bank project The Water Bank project is a recent rural housing project intended to house 40 families in Maharashtra. FAB LABs2 share core capabilities. The centre was founded in 1983 by the late Dr Kalbag. The complex is an experimental 8 million rupee investment (£116. highspeed embedded processors. Students undertake projects in numerous areas including water resource development.000 litre water supply yearround. 900 square foot work area. Maharashtra. where the engINdia team worked with Vigyan Ashram and the local community to gain an understanding and appreciation of the development issues concerning rural India which could be tackled through engineering. 1 2 http://www.youth. The kit includes: a computer controlled lasercutter. The complex is intended to provide each of the families with a guaranteed 2. to recognise patterns and to form and test hypotheses. workshop technology. a 300 square foot greenhouse. Massachusetts Institute of Technology (MIT) and the Indian Institute of Technology Bombay. Maharashtra. environment agriculture. engINdia1 is a partnership between 6 students from the University of Cambridge.edu/ 4 . 350 square foot cave house. An expedition was conducted during the summer of 2005 to the area of Pabal. Places are oﬀered both to students in India’s public high school system. The domes have soil packed over them which was never anticipated in Vigyan Ashram’s original design. 2. so that people and projects can be shared across them. to experiment. transport. a terrace.5 acres of land. construction. MIT have established a FAB LAB in Pabal through their “Bits and Atoms” program. Vigyan Ashram’s philosophy [Kalbag. 2004] places an emphasis of teaching science through practice. a large scale milling machine. a micron resolution precision milling machine and programming tools for lowcost.2.cba. Work that began in January 2005 has been halted due to concerns over the loading applied to a series of geodesic domes which form a large portion of the whole complex.engindia. a courtyard and 0. Through this relationship.
If appropriate. he inﬂuenced numerous architects and some engineers to a greater extent than many eminent members of their professions [Zung. An inhouse analysis method would allow assessment of the dome’s structural response in situations previously not considered. Though he never possessed formal qualiﬁcations in engineering or architecture. 2. This would be a further step to empower the locals to solve their own engineering problems in the future. it emerged that they wished to alter the original brief to focus the design of a bespoke structural analysis package. Poliedro de Caracas (Caracas. capable of encompassing maximum volume with minimum surface area.3. This would allow Vigyan Ashram to analyse the structural response of the domes inhouse without the need for buying costly ﬁnite element software licences or requesting outsider help.3 2. students at Vigyan Ashram would also use the package to further their understanding of structural behaviour. The earliest geodesic dome was designed and built in 1922 by Walter Bauersfeld in Jena.1 Geodesic domes Buckminster Fuller Engineers and architects have always held a special interest for structural systems that enable them to cover large spans with minimal interference from internal supports. that the geodesic dome became a vogue form of design. Vigyan Ashram hoped this would increase the number of potential uses of the structure and its market demand. He ﬁled for a patent in 1951 for an improved version of the geodesic dome design. It was not until the 1950s. and largely due to the eﬀorts of Buckminster Fuller. are one of the oldest structural forms and have been used in architecture since the earliest times. 2001].2. Vigyan Ashram’s original brief requested that an assessment of the loading encountered by the domes should be performed and design alterations should be recommended to accommodate the expected loading. It is perhaps no surprise then that dome structures. which has since been used in structures such as the Tacoma Dome (WA. USA). UK). Venezuela) and The Eden Project (Cornwall. USA. Vigyan Ashram and engINdia approached the student led charity Engineers Without Borders UK3 to request that more research be conducted into the design of the geodesic domes manufactured in Pabal.org/ 5 . Buckminster Fuller was an American born in Massachusetts.4 Structural analysis When the Water Bank project halted. 1979]. Following discussions with Vigyan Ashram.2. There is a clear need for a method to model the structural response of the geodesic or “Pabal” dome under various loading conditions. Germany [Makowski.ewbuk. 3 http://www.
Buckminster Fuller’s passion for geodesic structures came from the design’s aﬃnity with Nature. If the chords that join the vertices are straight lines rather than curves.polyhedra made up entirely of congruent regular polygons. 1981] Buckminster Fuller based his original design on the sphere division of an icosahedron. Fifteen great circles also fully deﬁne the primary bracing of a geodesic dome. (see Figure 1) although geodesic domes have since been constructed using octahedron and dodecahdreon symmetry systems to circumvent Buckminster Fuller’s patent. for Class I subdivision. B and C in Figure 2. when combined in the right conditions produce water. molecules. Similarly. To obtain a more regular network. an eﬀect he described as the “energeticsynergetic” geometry of his domes.2 Deﬁnition Figure 1: Isosahedron Figure 2: Great circles [Morgan. most eﬃcient forms of construction. The aforementioned shapes are all part of the family of platonic solids . a secondary bracing is introduced. the vertices of which may also be described by the intersection of three great circles (circles with a diameter equal to that of the sphere) and are referred to as geodesic points. planar triangles are formed and this creates the geodesic network commonly used in structures. such as A. 2. the dividing lines are parallel to the 6 . “Energetic” refers to Fuller’s belief that Nature always builds the most economic structures. Fuller deﬁned synergy as the “integrated behaviours of nature and the behaviour of a whole system unpredicted by the behaviour of any subassembly of its components. An icosahedron exploded onto the surface of a sphere produces twenty equilateral spherical triangles. but impractical to use in most circumstances as members quickly develop excessive slenderness ratios as the diameter of the dome increases. and this could not be predicted by the individual properties of either gas. modularly dividing each equilateral triangle into a number of subdivisions.” For example. would form the lightest.3. and crystals. The primary bracing is truly geodesic. hydrogen and oxygen gases. the geodesic dome exhibits a stiﬀness and rigidity greater than that predicted based on the sum of the individual components that make it. He claimed that geodesic domes built upon principles embodying force distributions similar to those of atoms. There are two possible classes of geodesic subdivision.
1984] edges of the primary bracing (Figure 3). Class I subdivision produces geometry where the edges of the triangle lie on a great circle. 7 . which is advantageous for fabrication. this may not be achieved with a Class II subdivision. Class II subdivisions require a smaller number of bar lengths. Odd order frequency subdivisions are nominally referred to with the suﬃx 3/8ths or 5/8ths. The frequency is often referred to in shorthand as a number. (a) V2 (b) V4 (c) V8 Figure 5: Example subdivision frequencies [Motro. as an equatorial perimeter ring is only produced for even order frequencies. 1984] Figure 4: Class II subdivision [Motro. Class II domes can only be achieved with an even frequency of subdivision. and this produces a less uniform stress distribution.Figure 3: Class I subdivision [Motro. the triangles are no longer perfectly equilateral . with the preﬁx “V”. 1984] A subdivision. however. in Class II. to indicate respectively if the ground ring is above or below the equator of the geodesic sphere. the diﬀerences between individual bar lengths are resultantly greater in a Class II dome. Odd order frequency domes cannot produce a hemispherical shape.the bars forming the skeleton show variations in length. and the number of diﬀerent lengths required to fabricate the dome increase with the frequency of subdivision. or “frequency” is deﬁned by the number of triangles each edge of the primary bracing is divided into (Figure 5). which leads to simple design of a hemisphere with planar connections. the dividing lines are perpendicular to the primary bracing (Figure 4). It should be noted that if secondary bracing is introduced. Additionally.
This is why in practice domes cannot be built from one uniform member size. quake resistant designs have evolved through trial and error exercises over the centuries. 2. i. a portmanteau of tensional integrity. they can not be ﬂattened into a plane without shrinking or stretching of the chords.. and a geodesic network produces a structural form with self stabilizing properties. Hawaii. where the curvature of any point is the same sign in all directions. While radial cable domes may exhibit greater stiﬀness to uniform loads. In high risk earthquake areas such as Japan.3 Explanation of behaviour The manner in which a braced dome behaves depends on the conﬁguration of the members.e. • The degree of compactness in a building correlates with its resistance against seismic shock. the rules listed below are extracts from a planning aid for architects and engineers [Dowrick. 2002]. where surfaces are of positive Gaussian curvature. Modern design guidelines have emerged from such empirical data. making the structure ideal for optimising the manufacture of components and prefabrication. Makowski [1965] points out that geodesic domes are exceptionally good in this respect.e. while compression forces are local and discontinuous. to convey the concept of coherence and resilient elasticity of geodesic networks. The 44. Morgan [1981] describes how the Kaiser Aluminium Company (USA) was one of the ﬁrst companies to take out a full license under Buckminster Fuller’s patent for a concert hall in Honolulu. The triangle is a planar ﬁgure which has maximum rigidity accomplished with the least eﬀort. Synclastic surfaces are not developable. such as geodesic domes have a high stiﬀness in all directions and are kinematically stable (no mechanisms) when idealised as a space truss. Tension forces are global and continuous. Richter [1975] argues that the use of geodesic subdivision produces structures of greater strength than conventional braced domes. and within 24 hours a concert was held at the venue by the Hawaiian Symphony Orchestra to an audience of 1832. i. Buckminster Fuller coined the term tensegrity. The forces in a geodesic network are an equilibrated combination of tension and compression.3. A dome which is not fully triangulated is not kinematically stable when idealised as a truss and stiﬀnesses may vary greatly in diﬀerent directions on the dome’s surface. Symmetrical triangular systems provide the most economic energy ﬂow.2.2m dome was erected within 22 hours of the components arriving.4 Strengths and weaknesses A dome is a typical example of a synclastic surface.3. 1978] and are used to highlight some of the properties that make the geodesic dome a highly earthquake resistant design. even for sizable spans only a small number of diﬀerent bars sizes are required. However. Fully triangulated domes. triangulated domes demonstrate greater stiﬀness to non uniform and concentrated loads [Kardysz et al. A hemispherical structure is the most geometrically possible compact form 8 .
The most common objections are functionalist. lighter and faster than cladding using masonry.21/m2 . Due to the light weight. and empirically has been proven to survive even severe earthquakes with only minor cracking in the cladding. without the need for specialised equipment. The cladding forms an insulating thermal mass that keeps the house cool all year round. located close to the building boundary and distributed evenly throughout the structure. unless custom made. leading to a perceived loss of privacy. enclosing maximum volume with minimum surface area. enclosing a large area without need for internal supports. and the generally uniform load distribution of geodesic dome structures. • Reinforcing elements should be rigid and symmetrically organised. costing only RS. • The centre of gravity should be as low as possible. especially through the core grid. Domes can be interconnected to form interesting housing designs and add further rooms or space. The kit requires no brick masonry. so is easy to transport. January 2009).200/f t2 (£9. The geodesic dome is. 2.3. not without disadvantages. A house kit may normally be assembled in one day using only simple nut and bolt construction. however. • The construction should be elastic and deformable to a certain extent. The Pabal dome is based on a V3 5/8ths Class I dome as shown in Figure 6. The dome is made of prefabricated equal angle struts and disc hub joints made from mild steel. deep foundations are not normally required. also presents a problem due to the curved walls of the structure which can result in some loss of space. the aesthetic appearance of the dome is largely dependent on how these closures are treated. Construction of shallow foundations allow a considerable saving in time and money over deep foundations.5 The Pabal dome Vigyan Ashram’s Pabal dome design is a developed version of the basic geodesic dome structure. Furniture. Geodesic domes produce extremely light skeleton structures that are very stiﬀ and rigid. the round shape of the dome perimeter. A dome has a lower mass point than any cuboid structure of similar proportions. eﬀective sound isolation is diﬃcult to achieve through partitioning of rooms. The basic structure is a cost eﬀective form of design.of construction. a form of spray on concrete that is cheaper. 9 . One complaint is that the perimeter chords following the shape of an icosahedron produce an irregular or ragged line that may be objectionable on architectural grounds. the geodesic skeleton does exactly this. The ﬁnished skeleton structure is clad using multiple layers of chicken wire reinforcing mesh and ferrocementing (or “guniting”) technology. The geodesic dome exhibits this behaviour. due to the hemispherical shape of the structure.
where joints are assumed continuous. for example. 2007]. Matlab and C++ based FEA programs were alternatives also considered.. The equivalent shell method may be divided further into two subgroups . but space frame and space truss idealisations are by far the most common. leading to a set of governing diﬀerential equations that are solved using a harmonic solution. computers have become more aﬀordable and more powerful. and the use of ﬁnite element analysis has become increasingly popular method of modelling structures that are too time consuming or complex to resolve by hand [Coates et al. where all joints are assumed pinned.4 2. it has to be represented by an idealised mathematical structure whose behaviour models suﬃciently closely that of the original structure. The second group of methods allow the analyst to tackle the discrete structure directly. as such software is widely available and signiﬁcantly cheaper than dedicated analysis packages. 1988]. Equivalent shell methods are best suited to early design work or for structures too large to be analysed discretely.Foundation closure Disc joint Door Figure 6: Pabal dome 2. Both methods use approximations which aim to smear the eﬀect of discrete members uniformly over the surface of an equivalent shell. Both discrete structure methods produce a large set of simultaneous equations that can only be feasibly solved with the aid of a computer. Over the years.4. Vigyan Ashram favoured the development of a spreadsheet based discrete structure method. and space frame analysis.orthotropic shell theory and ﬁnite diﬀerence theory. 4 10 . the equilibrium method discussed by Quintas and Avila [1993] or Endogenous force analysis [Burkhardt. Two idealised methods exist for braced dome structures: equivalent shell methods and discrete structure methods. Again. this may broadly be split into two subgroups4 : space truss analysis.1 Braced dome analysis Idealisations Mullord [1984] states that before any engineering structure can be analysed. but were ruled out due to the required presupposition of programming Other direct analysis methods do exist.
more technologically complex method was not necessarily the most relevant technology for Vigyan Ashram to use. Linear analysis. However. Nonlinear methods are generally more technical and more computationally exhaustive.4. Microsoft Excel 2003. Linear analysis assumes a linear stress strain relationship. it was important to understand the level of complexity of the diﬀerent methods. 2. a more accurate. However.knowledge in order to understand the workings of the structure stiﬀness method. instability eﬀects beyond the point at which the material yields require nonlinear analysis to account for member eﬀects such as plastic yielding. Figure 7: Brown University’s 2D FE program 11 . It is important to understand that Vigyan Ashram’s students will have limited programming knowledge due to the vocationalstyle training used there to teach. 1989]. being less computationally intensive and requiring less engineering knowledge to understand was therefore better suited for use in Excel. and are based on either the ﬁnite element method or on the beamcolumn approach. was mutually decided to be the most appropriate platform to develop the solution with. Nonlinear analysis methods have been extensively studied by many authors [Meek and Loganathan. Vigyan Ashram required a straightforward analysis method that produces sensible design information for a comprehensive range of loading options. and together with information about the material behaviour can be used to check for local member or joint failure.2 Analysis Discrete structure methods can be modelled using linear or nonlinear analysis. but produce a more accurate solution. a widely available spreadsheet package that Vigyan Ashram are familiar with using. To decide which analysis method was most appropriate.
see for example Stroud [2001]. resulting in a very time consuming and unnecessary data entry exercise. 3 Theory The following section introduces the underlying theory required to understand how the FEA package covered in Section 6 was developed. that is.requiring the user to specify each nodal coordinate. subtraction.2.4. but is not essential. Similarly. inversion and transposition6 . The emphasis here is on the processes involved in forming and solving a system. Teh and Morgan [2005] highlight the beneﬁts of the platform for teaching the stiﬀness method to ﬁnal year university students at Curtin University of Technology. understanding matrix addition.engin. member stiﬀnesses. a method of forming basic 2D pin jointed analysis developed by Brown University.3.brown. and this is internally carried. 3. The reader is assumed to have a grasp of matrix mathematics. only the method for a space truss system needs to be discussed here for the reasons described in Section 2. USA5 is available freeware for solving simple nodal applied loads (Figure 7). most maths textbooks cover the topic of matrices. It can deal with a wide range of design situations.1 Stiﬀness method The stiﬀness method allows us to analyse a structure which is an arbitrary assembly of simple structural members by breaking down the components into “elements” (members) and “nodes” (joints).edu/courses/en3/notes/Statics/Structure_tutorial/Structure_tutorial. To achieve this. 5 6 http://www. these examples are limited to two dimensional analysis of a limited number of members. causing a resultant change in length u. The basis of the stiﬀness method for a structure limited to pin jointed elements is that every member has analogous properties to those of a spring. a fresh approach and a bespoke design were required.htm If not. including space frame structures (where joints are treated as continuous).3. rather than justiﬁcation of why a particular method is most suitable (this is covered in Section 2. However.4). multiplication. ﬁxities and each nodal load.3 Previous work The concept of using Excel for the purpose of structural analysis is not entirely new. A previous understanding of the stiﬀness method will be helpful. The interfaces are also at a very basic skeleton level . Vigyan Ashram’s request was for a user friendly interface that would streamline the analysis process of a geodesic dome and work with a variety of diﬀerent load situations. 12 . an axial load F may be applied to the end of a spring. Some of these programs also make use of built in matrix manipulation functions that are not designed to work with large matrix problems. Australia. However.
u is a resultant change in length and L is the length of the original member.4. v and w (Figure 8). 7 σ = F/A FL = . for a linear elastic material. the principle needs to be extended to work for a three dimensional system. (1) It is possible to model the response of a structure by connecting together a system of individual spring elements and solving to determine the displacements. or a diﬀerent material is used. it is convenient to express the local stiﬀness of an element using matrix notation: F = The reader may wish to follow a worked example for a 1D or 2D structure stiﬀness method from a textbook. (3) L where F is the axial force in the member. as each node can displace in three dimensions: u. before reading the remainder of this section. k = EA . Though the derivation of element stiﬀness is quite elementary. Intuitively. or stiﬀness k (Equation 1). Consider the expressions for stress (σ) and strain ( ) in an element σ= and F A (2) u . (5) L Comparing Equation 5 to Equation 1. a way of determining the stiﬀness of each member must be found. these expressions may be related using Hooke’s Law: = E= and rearranging gives EA u. In order to ﬁnd a structure’s response to a known set of forces. however a total of six degrees of freedom (DOF) must now be considered for each element. expressed as F = ku . Now. it has been included to help explain the steps taken to post process the spreadsheet data in Section 6. the value of k will change. it is clear that an expression for the spring constant in terms of geometric and material properties of the material exists. L Now that the stiﬀness for an individual element has been determined. A is the area of the cross section. u/L uA (4) 13 . such as Mullord [1984].The relationship between applied load and the resulting displacement is a property of the element.1. known as a spring constant. Rather than dealing with multiple linear expressions. if the section is made larger.7 In 3D the expression relating forces and displacements is much the same.
[1988] and most good texts on FEA. [Kglobal ] = [T ]T [klocal ][T ] . in shorthand form {flocal } = [klocal ] {u} . a transformation matrix [T] is applied to the local stiﬀness matrix to obtain a common a global stiﬀness matrix for each element8 . 8 (8) (9) 14 . 0 R0 and the general form of R0 is The derivation of this equation has been omitted. z’ (with the x’ axis aligned with the direction of the truss element as shown in Figure 8). where R0 0 [T ] = . righthand rule coordinate system x’. but is covered in Coates et al. y’. In order to obtain meaningful stiﬀness matrix. the local Cartesian coordinate system is unlikely to coincide with the coordinate system of the global axis X. In a 3D structure. (7) Equation 7 uses a local.Figure 8: DOF for a pin jointed element in a local coordinate system f x 1 fy 1 f z1 f x 2 f y 2 f z2 = 0 0 EA − L 0 EA L 0 0 0 0 0 0 0 0 − EA L 0 0 0 0 EA 0 L 0 0 0 0 0 0 0 0 0 0 0 u1 v1 0 0 w1 0 u2 0 v2 0 w2 (6) or. Y. Z.
In FE programs. In the case of a pin jointed (truss) structure. Lx /L [R0 ] = −Lx Ly cos γ−LLz sin γ √ L L2 +L2 z x Lx Ly sin γ−LLz cos γ √ − Ly /L L L Lz /L √ L2 +L2 z x L2 +L2 z x √ L2 +L2 cos γ z x L L2 +L2 x z L sin γ −Ly Lz cos γ+LLx sin γ √ L L2 +L2 z x Ly Lz sin γ+LLx cos γ √ . L L2 +L2 x z (11) √−Lz 2 Lx +L2 z 0 √ Lx 2 Lx +L2 z Once every element stiﬀness matrix has been transformed into a global coordinate system. for a space truss. The angle γ is a rotation of the element about the x’ axis in Figure 8 to align the cross section of the element. For a truss. each element stiﬀness matrix is 6 × 6 and the ﬁnal structure stiﬀness matrix size [K] is determined by the number of nodes multiplied by the number of degrees of freedom per node. and its projected lengths on each of the global axes X. this is often achieved through the use of a steering matrix which links the local numbering scheme to the global numbering (see Figure 9). γ = 0 and Equation 10 simpliﬁes to Lx /L −Lx Ly [R0 ] = L√L2 +L2 x z L /L √ y2 2 Lx +Lz L Lz /L −Ly Lz √ . Y and Z are Lx . the structure stiﬀness matrix in Figure 9 would have 15 . (10) where L is the length of the member. Ly and Lz respectively. 1 1 2 3 4 5 6 1 2 4 4 5 6 7 8 9 5 6 7 8 9 3 4 5 2 3 4 5 6 Kelem1 6x6 1 2 3 4 5 [K] (12x12) 6 7 8 9 10 11 12 Kelem3 Kelem1 Kelem2 Kelem3 Kelem1 6x6 6 7 8 9 10 1 1 2 3 10 11 12 2 3 10 11 12 11 12 Kelem3 Kelem3 Kelem1 6x6 Figure 9: Assembly of the structure stiﬀness matrix from element stiﬀness matrices. Nodes that are shared between elements will have overlapping contributions to the same cells of the stiﬀness matrix. Hence. For example. it only remains to assemble all the matrices to form the structure stiﬀness matrix. the shape and orientation of the cross section is unimportant as bending moments are ignored.
The greatest width of any row in the matrix is termed the “bandwidth”. An eﬃcient numbering system will produce a narrowly banded set of values on the leading diagonal of the populated matrix. in terms of a known global force vector {F } and the structure stiﬀness matrix [K] {F } = [K] {U } . Solving Equation 12 is not as straightforward as it may seem. boundary conditions are applied to the problem to deﬁne it uniquely.to be composed of four nodes. This is normally achieved through ﬁxing certain degrees of freedom in the structure to represent anchorage to the ground. a matrix version of Equation 1 has been developed. Finally. where [K][K]−1 = [I] . First. as division is not possible in matrix algebra. the matrix equivalent of unity. This allows us to then solve Equation 12 as 16 . bandwidth bandwidth bandwidth bandwidth (a) Low bandwidth (b) High bandwidth Figure 10: Structure stiﬀness matrix bandwidth The node numbering system used in the structure is often critical to the eﬃciency of the solver.2 Solution methods In order to gain a useful output from Equation 12. A less eﬃcient numbering system produces a much wider spread of results symmetrically about the leading diagonal and therefore produces a larger bandwidth (see Figure 10). (12) 3. (13) and [I] is an identity matrix. especially for large structural problems with thousands of elements. as 4 × 3DOF produces the 12 × 12 matrix shown. a method of solving it is required. The closest equivalent matrix operation that emulates division is to ﬁnd its inverse [K]−1 .
(14) However. stresses and forces using Equations 2 and 3. First. considerably more computationally eﬃcient methods are used. the process of inversion is computationally intensive and slow. • Direct solvers. Rather than tackle the inversion directly. These solvers use methods such as Gaussian elimination and LU decomposition and are guarenteed to arrive at a solution if the matrix is nonsingular. A positive change in length indicates the element is in tension. according to Kreyszig [2006]. it is a relatively straightforward process to obtain the internal axial forces that form an equilibrium with the externally applied force vector.2. (15) The diﬀerence between the x’ directional nodal displacements u2 − u1 (see Figure 8) give the overall change in length of the element which is then used to calculate the strains. Many of these methods may be found in mathematical textbooks9 . These solvers. the solver quickly becomes a dominant part of the run time. As Gaussian elimination is a particularly well documented way of solving linear systems. such as the conjugate gradient method. Direct solvers also suﬀer from an accumulation of roundoﬀ errors. a negative change indicates compression. Trevelyan [2007] categorises solvers into two groups. it is unnecessary to reiterate the method in great detail here.3 Post processing Once the displacement vector in Equation 12 has been found using a method described in Section 3. make an initial guess and progressively iterate towards a solution until a convergence tolerance is reached. 9 See for example Kreyszig [2006]. as then the system may easily be solved by back substitution. meaning that for large matrix FE models. The standard method. based on a systematic process of elimination in the rows of the matrix in order to reduce the system to triangular form. The number of ﬂoating point operations required to solve a system using a direct solver increases with the cube of the number of equations. Gaussian elimination has variants such as GaussJordan elimination which has its own advantages and drawbacks. however only Gaussian methods will be mentioned in more detail here. 17 . for a large matrix. 3.{u} = [K]−1 {F } . • Iterative solvers. is a popular way of solving linear systems. the transformation matrix from Equation 9 has to be applied to the global displacement vector in order to revert the displacements to the local coordinate system: {ulocal } = [T ] {U } .
γsat is the bulk unit weight of the soil and z is the depth below some datum line at the surface of the soil.3. Rankine found that the relationship between the horizontal and vertical stresses in the active region10 of the soil were related by an active pressure coeﬃcient Ka . z. For a horizontal soil mass with the water table well below the point of interest. and hence eﬀective and total vertical stresses at a depth below the surface. (16) where σv is the total normal (vertical) stress on the soil mass. 2004]. σv is the eﬀective normal stress. Strand is a general purpose ﬁnite element analysis system. controlled by the soil’s intrinsic shear strength parameter φ: σh = Ka σv .1 Geodesic dome analysis Introduction to Strand7 A commercial FEA package called Strand711 was used to build up an initial geodesic dome model. which considers the state of stress in a soil mass when the condition of plastic equillibrium has been reached. the retaining structure is being pushed against a resisting soil mass. when shear failure is on the point of occurring throughout the mass. are equal and equivalent to the weight of all material per unit area above that depth: σv = σv = γsat z . i. In an active pressure zone.e. the retaining wall is pushed away from the soil mass in response to the action of the soil.strand7.3. Rankine’s theory of earth pressure is used. the pore water pressure is zero.3. 1 + sin φ (18) (17) 4 4. where Ka = 1 − sin φ . According to Rankine.4 Lateral earth pressure The importance of the forces transmitted through the soil skeleton from particle to particle was recognised in 1923 when Terzaghi presented the principle of eﬀective stress.com/ 18 . To relate the distribution of vertical stresses to those induced horizontally against a retaining structure. which applies only to fully saturated soils [Craig. In the passive zone. http://www. wheras horizontal stresses increase or decrease depending on the local movements of the soil. vertical stresses in the soil remain at a constant geostatic value. used in a wide range 10 11 The passive case calculation has been omitted for the reasons discussed in Section 4.
However. and element stiﬀnesses are described locally and globally by Cartesian coordinates. assuming the model is deﬁned correctly. as the stiﬀness method discussed in Section 3. It is a system particularly suited to generating spherical geometries.2 Coordinate generation The mathematical determination of geodesic dome coordinates may be achieved in one of two systems: Cartesian (presented by Davis [2007]) or polar (presented by Kenner [2003]). it is an eﬃcient tool for modelling the behaviour of structures. to a competent user with a good understanding of the theory behind the program. the forces that were considered to act on the structure. without showing the intermediate calculation steps. 4. and it operates a black box approach to analysis. While a powerful analysis tool. It therefore is not suited to teaching or explaining the FE method. civil. using algorithms based on formex algebra (see SanchezAlvarez [1984]). The package was chosen because it was the most appropriate general purpose structural analysis software available at Durham University’s School of Engineering. However. a righthandrule Cartesian based coordinate system was used for developing the Strand model. aeronautical and biomedical engineering.1 is based upon Cartesian transformations. It was hence logical that the basis of the model should be deﬁned using a global Cartesian system (Figure 11). and the results that it produced under various analyses. All forces and displacements in FEA are expressed in Cartesian form. Strand will then produce a solution. Figure 11: Global coordinate system used for analysis A polar coordinate system deﬁnes the position of a coordinate relative to the origin at the centre of a sphere by a radius r and two angles θ and ψ. licensing costs for this software are high. structural. The user is required to deﬁne a model and run the solver without guidance or explanation. 19 .of applications in mechanical. The following sections describe the generation of geodesic geometry in Strand. Strand would not itself be an appropriate piece of software for Vigyan Ashram.
2007].4). 4. 4. the internal ﬂoor area produced may be of greater concern than the overall external dimensions of the structure.1 Seismic and impact loads It has already been demonstrated that the Pabal dome is an earthquake resistant form of design (Section 2. In addition to generating coordinate data. an odd order frequency dome does not produce an equatorial perimeter ring because the structure is not an exact hemisphere.3. When designing a structure.cadreanalytic. as the Pabal Dome’s assembly guide initiates the dome construction by marking the desired dome radius at perimeter ring (ﬂoor) level. This was indicated to be the case for Vigyan Ashram. 4. wind. This factor was implemented as an option in the coordinate generation spreadsheet. Normally.htm 20 . impact damage was considered very unlikely given the residential use of the structure and that compound walls are often built around Indian housing [Rolly. but. as discussed in Section 2. As it was not therefore necessary to consider these actions for normal design situations. the Maharashtra region of India maintains a temperature of 2030◦ C all year. The method that nodal forces were calculated from the applied loads is also described here.. CADRE is a design utility for generating a wide variety of geodesic and spherical 3D wire frame and surface models for Computer Aided Design (CAD) and FEA applications.2 Snow Snow loading is often critical to dome design [Knebel et al. Setting the radius of a V3 5/8ths dome to unity in CADRE produced a set of coordinates which could be scaled to deﬁne the geometry of the dome for any radius in a spreadsheet. soil. it was assumed dynamic load eﬀects were beyond the scope of this project and could be neglected from the analysis. 2002]. the radius is deﬁned at the equator of a sphere.2.3 Loading This section discusses the diﬀerent loads that could act upon a geodesic dome structure. snow.3. The totality of possible actions on the structure were considered to be: self weight.012 was used to generate the geodesic geometry for the Strand model. and has a record 12 http://www.3.3.A trial version of a package named CADRE Geo 6. saving a considerable amount of time on assembling the elements connecting nodes manually. However. imposed.com/cadregeo. seismic and impact. CADRE presented an option to expand the perimeter ring of the dome to equal the radius of the structure by applying a scaling factor to the coordinates (found to be 1.016). Additionally. the package was able to produce dfx ﬁles that could be imported directly into Strand. explaining which were considered important and why.
This is formed from the subaqueous decomposition or in situ weathering of basalt rocks. 2003]. It was assumed that even during the monsoon aquifers would not recover to a level above the foundations.3. which was a reasonable assumption to make as the region is prone to drought [Infobase. 4. This meant that underground aquifers were likely to be relatively deep below ground level. The surface area of the dome was then divided between the nodes. 2006].3 Soil Datum Line Z X σv σh z σv = γz z σh = KAγz Figure 12: Horizontal and vertical stresses acting on a subterranean dome (after Landry [2002]) Soil loading was a necessary consideration because of the dome’s subterranean use in the Water Bank project (see Section 2. Rankine’s earth pressure theory (covered in Section 3. It was assumed that the loaded dome would act as a retaining structure against any soil piled against it. as shown by the regional soil map in Figure 13 [Shroﬀ and Shah.4. The predominant surface deposit of the region surrounding Pabal village is Black Cotton soil. It was therefore safe to assume that snow loading would not be a design issue and could also neglected from the analysis.coldest day of 10◦ C [MapXL. The design assumes that the water table is below the foundations of the dome.4) was applied to calculate the total horizontal and vertical stresses induced by the weight of the soil. so that pore water pressure would not make a contribution to the stress state of the soil as described in Section 3. based on a depth z below a user deﬁned datum soil level. In order to calculate the forces upon the structure. The stresses were calculated for an active pressure case. 2000]. with no resistive mass of soil acting passively.2. allowing the calculated stresses to be converted into nodal forces. which 21 . It was conservatively assumed that the soil loading would be carried by the dome framework alone.3). as this reduced the complexity of calculating horizontal pressures for asymmetric loads. as shown in Figure 12.
Montmorillonite is a dark swelling clay commonly known in India as Black Cotton soil.Figure 13: Regional soil deposits of India (after Shroﬀ and Shah [2003]) produces the mineral Montmorillonite in an alkaline environment. They state that cyclonic storms do not normally extend further than 60km from the coast and that hurricane actions normally occur in the north east of India. In actual fact. 4. they include factors of adjustment that account for cyclonic wind speeds).4 Wind Bhandari et al. In later work developing the Excel spreadsheet. 2000]. The basic wind speed and design wind pressure were determined to Indian Standards using IS875:Part 3 [Bhandari et al. which would act to reduce load on individual elements (see Figure 14). especially during the monsoon season. These were chosen in preference to Eurocodes as they are more representative of Indian wind conditions (for example. the external pressure proﬁle of a spherical structure is partly negative (i. Stabilising wind pressures produced at the back of the dome due to vorticies were ignored. more soil types were added as options for analysis. This places Pabal (which is around 100km from the west coast) in a relatively low risk area. the calculations and assumptions behind this value may be found in Appendix C. [2003] give general guidance to the treatment of cyclonic storms and hurricanes. due to its prosperity for growing cotton [Ranjan and Rao.3. 1987]. It was conservatively assumed that the wind pressure acts only from one direction and positively.584kN/m2 . 22 . which are characteristic of India. The load contribution due to the soil in the analysis (Section 4.5) assumed that the dome was buried in Black Cotton soil. suction) for angles between approximately 45◦ and 135◦ relative to the wind direction..e. The design wind pressure pd for the Pabal region was determined to be 0.
so that results would be directly comparable.3.Figure 14: External wind pressure coeﬃcients for a spherical structure (from Table 18. this also allowed the dome to be considered in situations 23 . which in turn was a function of the dome geometry.5 Self weight and imposed loads The main components that contribute to the weight of the structure are the elements and the ferrocrete shell. Table 20 of IS875 gave the force coeﬃcient as Cf = 0. but was discarded in favour of a spreadsheet based equivalent for two reasons.7. The eﬀective area was taken to be projected area of the dome facing the wind. rather. These were initially represented using truss elements and plate elements with speciﬁed densities in conjunction with Strand’s gravity function. (19) For a rough spherical structure. ﬁrst. Bhandari et al. FE analysis of plate elements is a considerably more complex procedure than the analysis of truss elements. a force coeﬃcient based on the geometry of the structure was used. The total wind loading on the face of the dome (found with Equation 19) was proportionally divided between the nodes by projected area to obtain the forces on individual nodes. to ensure the method of generating forces would be common for Strand and Excel. This arrangement was convenient for use in Strand. Though this may appear beneﬁcial. It was desirable to keep the analysis simple and focused on the geodesic skeleton. forces on individual members were not considered. In order to get the total wind load acting on a particular building. 4. and second. because including plate elements in the analysis adds to the rigidity of the dome. This automatically calculated the nodal force contribution when an analysis was run. [1987]) For simplicity. Aside from making the spreadsheet stiﬀness solver easier to follow. Together with the eﬀective face area Ae and design wind pressure pd this gave Fw = Cf Ae pd . forces were considered to act on the structure as a whole.
B and C make up the possible triangle sizes for any given geometry. A similar. and hence the total applied force on this node due to the shell would be 1 1 × ( AB) × ρshell g t . (21) 2 where Aelem is the cross sectional area of the members. a third of the load on each plate was allocated to each node.2. Once these equivalent support reactions for all nodes were calculated. then simply half the length of each connecting member is supported by the node under inspection. Figure 15: Example area division for a node For example.3). The area 1 of each plate in this case is 2 AB. (20) 3 2 where t is the thickness of the shell.where it may not be built with a reinforced concrete shell. the node under inspection is bordered by ﬁve triangulated sets of elements. each consisting of two element lengths. but simpler method was applied to ﬁnd the distribution of member forces between nodes. hence each triangle was bisected to evenly split the plate load distribution between the nodes. in Figure 15. two of A and one of B. for example. three unique chord lengths A. assuming element homogeneity. as each triangle was supported equally by three corner nodes. the area supported by each node had to be calculated. when providing the framework for the greenhouses mentioned in the Water Bank project complex (see Section 2. All 5 members connected to the inspected node are of length A in this example. The shell was assumed to be approximated by loaded triangular plates. It was therefore possible to determine the area of each triangle. the sum of reactions in Strand for an isolated self weight load case was compared to those determined Felem = 5 × 24 . In order to divide the self weight of the structure between the nodes in an equivalent way to Strand’s gravity method. hence the nodal applied force due to the elements’ self weight would be Fshell = 5 × 1 × (A × Aelem ) . and. ρshell is its density and g is acceleration due to gravity. For the V3 5/8ths dome.
which would consider behaviour under time varying conditions. • The door (see Fig.3.5 was ignored. joints and bearings. were ignored. time related eﬀects on the structure. the following assumptions were made for the analysis. so that no stresses were induced due to mismatched lengths. Table NA2.e. 1. such as earthquakes). EC030. • The components that make up the structure were homogenous and isotropic. • The structure’s working life was assumed to be 50 years (category 4. members. Section 5.in Excel to conﬁrm that they coincided. • The boundary conditions assume that the structure was fully ﬁxed at all its nodes at ground level. More components of self weight. Class I. i. • Members were assumed to be their correct mathematical geodesic lengths. • The analysis was static (where conditions are independent of time. 25 . • The good practice guidelines outlined in Eurocode 3. 6) and any windows in the dome structure were suﬃciently spaced and installed in such a way that they would not reduce the structure’s rigidity.1). such as creep. 2.1 apply. 3. Icosahedron based Pabal dome (with 61 nodes and 165 elements). including mesh density and joint weight were later included using the same distribution method to allow a greater amount of ﬂexibility designing the structure. as opposed to dynamic. The calculation model and basic assumptions for the calculations shall reﬂect the structural behaviour at the relevant limit state with appropriate accuracy and reﬂect the anticipated type of behaviour of the cross sections. 4. • The analysis was not rate dependent. • Any beneﬁcial contribution to the structure due to a stressed skin eﬀect from the reinforced ferrocrete shell described in Section 2. The method of analysis shall be consistent with the design assumptions. and the assumed building use was residential (category A). The imposed loading cases were similarly calculated.4 Assumptions In order to analyse the V3 5/8ths. and based on some of the considerations discussed previously regarding modelling forces on the structure. Analysis shall be based upon calculation models of the structure that are appropriate for the limit state under consideration.
4. which for a dome with a radius of 3m is a depth of 3. a choice reﬂecting the most common member type used by Vigyan Ashram. a coarse (gauge19) square welded mesh and three layers of ﬁner (gauge18) 1” chicken wire mesh14 . 14 13 26 . 15 Steel data from Cobb [2007]. for the reasons discussed in Section 2. 16 Recall that odd order frequency domes are not hemispherical (Section 2. described below. • The analysis was isothermal (independent of temperature). The purpose of the analysis was to explore which load cases were the most signiﬁcant and to provide some solutions to which the Excel based FE method could be compared to for accuracy. bedded on two grades of reinforcing mesh. 4. with its base expanded to equal the radius (see Section 4. Five load cases. The properties of these structural components are summarised in Table 1.5. The joints and elements were assumed to be fabricated from mild steel.3.2). A nominal mass of 1kg was speciﬁed for the joints.3. This was a reasonable assumption to make.2). Ferrocrete associated data from Hartog [1984].1 Load Case I: Self weight Based on the technical drawings contained in Rolly [2007] and the information provided by Vigyan Ashram.4.5. The ﬁrst assumed the soil would be buried uniformly around the dome up to its tip. using the recommendations for safety factors supplied by the British Standards Institution (BSI) Eurocode13 0: Basis of Structural Design [BSI. The dome was set to a radius of 3m. assuming that all joints were pinned.59m16 . The second considered a case of nonuniform burial. • The members were assumed to be suﬃciently squat so that buckling eﬀects could be neglected. were considered individually and then in combination.3. the thickness of the ferrocrete shell was estimated to be 50mm. 2007].• The structure was modelled as a space truss. and hence the height of the dome is not equal to the radius. 4. and the truss elements were assumed to be 25 × 25 × 3 equal angle sections made from rolled steel15 .2 Load Cases II & III: Soil Two soil load cases were investigated (Figure 16). a more realistic situation where only half the dome was buried under a soil mass.5 Analysis The dome analysis was completed in Strand based on the assumptions listed in Section 4. as this is one of the most common kit conﬁgurations used at Vigyan Ashram. Eurocodes will henceforth be abbreviated to “EC” and will always refer to BSI [2007].
22kN .Table 1: Element and joint properties Material Parameter Symbol Value Unit Steel unit weight γsteel 77. with the parameters17 summarised in Table 2. the face wind force to be applied was Fw = 0.3.93 kg/m2 Joint mass mj 1 kg (a) Load Case II: full burial (b) Load Case III: half burial Figure 16: Soil loads considered In both cases the soil used was assumed to be the regional topsoil: Black Cotton. 27 . 4. The prevailing wind direction was assumed to coincide with the global X direction (Figure 11).4.5. 8 (22) Data based on the information provided by Ranjan and Rao [2000].15 kg/m2 Mass of ﬁne mesh mf 0.2 kN/m3 Young’s modulus E 200 GPa Element cross section Aelem 157 mm2 Ferrocement unit weight γshell 22 kN/m3 Thickness of ferrocement tshell 50 mm Mass of coarse mesh mc 1.584 = 7.3 Load Case IV: Wind The wind load case on the structure was calculated following the assumptions described in Section 4. Using Equation 19.7 × 17 5 2 π3 × 0.
based on the safety factor recommendations for permanent and temporary loads on structures set out in EC0. and imposed loading was factored as a primary case temporary load. Ka 0. Combination II: This modelled a more realistic case of subterranean usage.35(Case I + Case III) + 0.6kN/m2 was allowed for. Combination I: This was chosen to model a simpliﬁed case of full subterranean usage of the dome. In the case of the maximum displacements. the global axis direction in which the displacement acted is speciﬁed after the number. wind on the exposed face of the dome a secondary temporary load.75(Case IV) + 1.75(Case V) 4. 4.05(Case IV) + 0. Combination II = 1.5 Combination load cases Three combination cases were considered.5. where the dome may be thought of as buried in part of an embankment.05(Case V) 2.5. This was a sensible value for roof maintenance access.35(Case I) + 1. Combination III: This model reﬂected a more general usage of the dome structure.455 4.4 Load Case V: Imposed A vertical imposed load of 0. as the dome would not be exposed to any wind. Combination I = 1.5. 1. 28 . Wind loading was ignored. allowing for two workmen with tools to work on the roof.Table 2: Black Cotton soil properties Parameter Symbol Value Unit Shear angle φ 22 degrees Bulk unit weight γbulk 20 kN/m3 Active pressure coeﬀ. Imposed loading was assumed to be the primary temporary load.35(Case I + Case II) + 1. Workers accessing the roof would be a rare occurrence. Combination III = 1. so imposed loading was factored as a secondary temporary load. Since the structure would be fully exposed to the elements. us . it was decided that wind would be the primary temporary load. The full soil loading and self weight cases were factored as permanent loads.6 Results The results of the Strand analysis are shown in Table 3.05(Case V) 3. The case considers the only permanent load to be the self weight of the structure.
7234 6. Perhaps not so intuitive was that asymmetric loads produced much larger member forces. II and V). An intuitive observation was that a symmetrical load (such as in Cases.1608 1.5. Similarly.0366 0.3014 69.2330 3.2248 (X) Case V 2. This supports the ﬁndings of Pakandam and Sarshar [1993]. This makes logical sense. Firstly.4470 7.9609 (Z) Combination II 39. which indicates that premature failure of members and joints could be expected.6312 143.9120 13.8199 (X) Case IV 1. the geodesic structures were found to deﬂect very little.4571 (Z) Combination I 21. I. Even under Combination III. (mm) Tensile Compressive us Case I 4.3680 102. as large external forces are acting to squash the dome inward.8791 (Z) Case III 33. To validate the structural stability of a postanalysis dome. produced a symmetrical distribution of forces in the members.Table 3: Strand analysis results Case Max.0611 1. It was also observed that the magnitude of the compressive forces exceed those of tensile forces predominantly in the soil load cases.1 Joint testing Introduction This section of the report describes how a relationship between the discjoint properties and the ultimate failure load of a joint was developed.7 Discussion Several conclusions can be drawn from this basic analysis. 5 5. yet it is the asymmetric load case that develops considerably larger axial forces. which would be expected to induce larger compressive forces inside the geodesic network.5681 9.1811 0.8374 99. the compressive forces in the members are considerably larger.7368 5.5500 0. Axial Forces (kN) Max. Disp. Under normal design conditions.0223 (Y) Combination III 7. the factored deﬂections and axial forces remain low for such a sizable structure. there is a clear indication as to why the geodesic dome structures are failing in the Water Bank project.5109 (Z) 4. it was necessary to know 29 . the soil loading cases produce displacements an order of magnitude greater than for the other cases.8961 (Z) Case II 10.9753 1. where a similar analysis was carried out using snow loads. This can be seen clearly from comparing Cases II and III: the former is buried with twice the volume of soil of the latter.
as shown in Figure 17. identifying dimensionless groups of parameters that required further investigation (the beneﬁts of this are discussed in Section 5. • Computational models of the discs were developed. In order to develop a parameter linked relationship.2 Pabal disc joints Figure 17: Close up of Pabal dome joint The joints used by Vigyan Ashram are stamped discs or “hubs” which have an array of predrilled holes to which the truss elements of the structure are bolted to. 18 30 . however these can eﬀetively be produced by using only four of the six connections on a hexagonal disc. 5. testing a large number of conﬁgurations using stress analysis in Strand. Two18 diﬀerent types of joint are required to build any geodesic dome conﬁguration .those that connect six elements (“hexagonal” discs) and those that connect ﬁve elements (“pentagonal” discs).whether the axial forces calculated would lead to failure of the joint. Arguably fournoded connections are needed for the ground level perimeter ring. a dimensional analysis was performed. These parameters were then investigated using both physical and computational testing: • Some joint replicas were fabricated inhouse at Durham University and tested to destruction to produce some “real life” data. The connections used by Vigyan Ashram are discs rather than lap joints. which meant that the Eurocodes for connection design did not oﬀer a method of estimating their capacity. as the code only provides a relationship for joints with a constant cross section.3).
19 This is assuming that instability (buckling) eﬀects in compression may be ignored. For the most common design conﬁgurations. and experiments are usually necessary to determine fully the way in which one variable depends on others. tension forces are usually the prevalent axial force in the structure (cf. as shown in Figure 18.The discs are fabricated from scrap mild steel. By inspection. which add bending strength to the disc and also aid with stacking for storage and transport. This simpliﬁes our task of developing an equation to predict the failure load of the disc joints. This means that even in cases where the largest magnitude axial force is compressive. as the gross cross section of the material resists the load. whereas in tension only the net area of the material resists the load (as reductions must be made for bolt holes). using a ﬂy press which stamps them into the correct shape. (a) Hexagonal disc (b) Pentagonal disc Figure 18: Uniaxial loading of bolt array Each individual bolt array is loaded uniaxially. 31 . Combination III in Table 3). including two distinctive crimps. if evidence can be provided to support that a joint will not fail under that magnitude of force in tension. as tensile failure alone can be conservatively considered. 5. a joint will always have a greater capacity against compressive failure than tensile failure19 . then neither will it fail in compression. Dimensional analysis is a technique that can be applied to reduce the number of experiments required to obtain these relationships.3 Dimensional analysis Massey [1998] states that complete solutions to engineering problems can seldom be obtained using analytical methods alone.
Young’s modulus was neglected as this relates to the failure strength of the material. Dependence on the length dimension [L] was then removed by dividing through by r. No quantity that may have aﬀected the problem was overlooked. all dimensions had been eliminated from Table 6. r/r gives unity and was hence eliminated. Table 4: Dimensional analysis: step 1 Quantity Symbol Dimensional forumula [M] [L] [T] Failure strength fu 1 1 2 Plate thickness t 0 1 0 Hub radius r 0 1 0 Axial force Faxial 1 1 2 The objective of dimensional analysis was to assemble the quantities into a smaller number of dimensionless groups. the variables were arranged in a way that removed their dependence on individual reference magnitudes. In this case. giving Table 6. leaving two reduced dimensionless groups: Faxial fu r 2 20 and t r (23) In Table 4. length [L] and time [T]) in Table 4. giving Table 5. it was eliminated. which is not a variable. Table 6: Dimensional analysis: step 3 [M] [L] [T] t/r 0 11=0 0 r/r 0 11=0 0 2 Faxial / (fu × r ) 0 2(1 × 2)=0 0 At this point. Table 5: Dimensional analysis: step 2 [M] [L] [T] fu /fu 11=0 1(1)=0 2(2)0 t 0 1 0 r 0 1 0 Faxial /fu (11)=0 1(1)=2 (2)(2)=0 As fu /fu gave unity. Similarly. The quantities considered were broken down into dimensional components (mass [M].The ﬁrst task in the process of dimensional analysis was to decide which variables may eﬀect the solution and express them in terms of their dimensional formula. 32 . temperature eﬀects were ignored as these were material dependent. one after the other. save for quantities that only had a indirect contribution20 . by dividing through by fu . dependence on the mass dimension [M] was removed ﬁrst. To achieve this.
A total of six discs were made: two pairs of 1. Though not an ideal representation. Engineering drawings of the joints were subsequently drafted (see Appendix D).4. measuring force/displacment data electronically into a Windows based software package.4.0mm hexagonal and pentagonal discs.5mm hexagonal and pentagonal discs and one pair of 2. hence a custom built press was instead used to fabricate joints from sheets of aluminium.In order to determine the relationship between these two groups. based upon the technical illustrations provided by Rolly [2007] and advice of Vigyan Ashram. this was however ruled out due to expensive shipping costs.1 Tension testing Joint fabrication Vigyan Ashram initially planned to send some sample joints for testing at Durham University. 5.4 5. 5.2 Apparatus Upper jaw Position dial Oil gauge Lower jaw Fine‐tune oil strain 50mm Oil pump control Figure 19: Denison loading machine A Denison tension/compression loading machine was used for the tests (Figure 19). At the time of use the machine was calibrated to a level complying with the United Kingdom Accreditation Standard (UKAS). The joint fabrication was a labour intensive process. This machine was capable of applying a uniaxial tension force of up to 25kN. It was therefore necessary to design and fabricate some replica joints. the results would still be suitable for comparison between practical and computational models. Durham University did not possess a press suﬃciently powerful to create the bespoke joints from mild steel. and hence only a limited number of discs were made. 33 . some experimental work was required.
which began to separate the top and bottom clamps hydraulically. whereas the Denison could only emulate one dimensional loading. As aluminium is a ductile metal. A geodesic tensegrity network of the dome is three dimensional. 5. an undesirable eﬀect. and the oil feed activated to around half load. 5.Finding a method of clamping the discs that would accurately represent the conditions to which the joint experiences in situ was problematic. It was decided to opt for a simple test rig using thumbscrews to attach the joint to predrilled (5mm thick) steel bars considerably stiﬀer than the test joint.4. nonexplosive failures were expected. The Risk Assessment and COSSH form for the experimental work may be found in Appendix F. but were hindered by the 50mm clearance between the teeth of the top and bottom clamps. as shown in Figure 20.4 Method statement The following operating procedure was used: 1. A joint was then selected for testing. The Denison was switched on. (a) Hexagonal rig (b) Pentagonal rig Figure 20: Test rigs for joints Several clamping methods were considered for the disc that would preserve the uniaxial tension case.3 Risk assessment The experiment procedure was considered fairly low risk and only standard operating precautions were employed. and its thickness checked and noted at ﬁve approximately even spaces around its perimeter using a micrometer (accurate to 34 . 2. but was the most feasible option to allow experimentation to proceed. The oil feed rate was reduced to hold the Dension stationary when the top and bottom clamps were spaced approximately at the right height to ﬁt in the test rig.4. This setup introduced a bending component into the tests.
When the disc failed. 21 35 .2kN was applied before releasing the clamp. The Denison datalogging software was loaded and the current position and registered force were set as datum values. The joint was attached to the test rig using thumbscrew connections and clamped in the top and bottom jaws of the Denison. 5. 5. 6.025mm/s throughout the test to maintain a smooth loading curve. 3. A load of approximately 0.4.005mm). and the positions of the top and bottom clamps were reset to an appropriate distance to repeat the above procedure for the next joint.±0. 7. the teeth would slip free unless physically restrained from doing so. The oil ﬂow was continually adjusted to keep the rate of separation at approximately 0. the test was stopped.5 Results (b) (c) (a) (a) (b) (c) Figure 21: Tension test results This was because the clamps are held together by the friction of the teeth. While continuing to hold the clamp closed with the winding key21 . 4. When little or no load was applied. the oil ﬂow was increased so that the teeth continued to separate slowly.
6 Discussion Several observations may be made about Figure 21. both at a 3% mesh density. preventing any translation or rotation. 5. An incremental force was applied in the plane of the disc using a face pressure at each of the two bolt holes.5 Stress analysis A series of nonlinear joint analyses were run using Strand to estimate the force at which joints would fail for various combinations of material parameters. calculating the resulting stresses and displacements in the material.1 Model development CAD models of the joints were created in Solidworks and imported to Strand. The plates were meshed for analysis in two stages: a tet6 element surface mesh.5. Each test applied a series of incremental loads to the plate. Before the investigation commenced. the mesh quality was checked using a convergence test. Three progressively ﬁner meshes were run to calculate the planar displacement of an element at the center of 36 . the Tshaped clamping arrangement (Figure 20(b)) for the pentagonal disc was fairly far removed from the uniaxial tension case that was of interest and this may have adversely inﬂuenced its load capacity. the pentagonal bolt array will have a greater equivalent area to resist loading. 5. The forcedisplacement plots for the remaining four joints are shown Figure 21. corresponding to the point where the disc deforms to align with the axial tension. In particular. so failure would be expected to occur hexagonal disc ﬁrst. Slight recoveries in strength were observed at point (c) as the joint crimps formed a necking region for torn material which temporarily impedes the shearing action. 5. Problem symmetry was taken advantage of to reduce the analysis to the loaded area of a single bolt array (as shown by the hatched area of Figure 18). in order to gain familiarity with the machine and experimental procedure before the full tests were run. save for a kink at point (a). followed by a tet10 brick mesh. representing the axial force transferred to the joint through the connection bolts. Each disc reaches a peak capacity (b) after which the joint begins to tear in shear. save for radial slip. This diﬀerence may be due to the less than ideal simpliﬁcations made for this test. This is counter intuitive as Figure 18 clearly shows that for a common load.5mm thick joints were tested as a trial run. Of particular interest was that the pentagonal joints in each case failed at a lower force than the hexagonal equivalents.4.The ﬁrst pair of 1. Each test plot begins approximately linearly. Boundary conditions were prescribed along each of the two symmetry edges of the element.
4.0mm) and three diﬀerent material properties (6082:T6 aluminium. However. Three diﬀerent plate thicknesses (1.0186015 1% mesh 718 0. S275 steel and S450 steel) were investigated. the disc radius was ﬁxed to 77. as the results were clearly converging on a solution at ≈ 0.2) was instead The total required number of analyses to investigate a relationship increases to the power of the number of variables. 22 37 . the disc radius should also have been varied to fully investigate the dimensionless groups in Equation 23.0187712 the outer bolt hole (Figure 22).(a) 3% mesh (b) 2% mesh (c) 1% mesh Figure 22: Mesh convergence Table 7: Convergence test results Mesh grade Run time(s) Displacement (mm) 3% mesh 6 0. The results (Table 7) conﬁrmed that the mesh quality was reasonable. Results were repeated for both hexagonal and pentagonal material conﬁgurations. an unfeasible amount of work given the time frame of this project22 . 2.5mm.0183190 2% mesh 18 0. (a) Aluminium (AL6082) (b) Steel (S275) (c) Steel (S450) Figure 23: Sketches of material stressstrain relationships A linear static analysis could not be used to predict the failure load. as this assumes the material always behaves linear elastically. a total of 54 analyses would have been required to investigate 3 radii variations. For duration of the computational work.0mm and 3. giving a total of 18 separate analyses.5mm. Nonlinear analysis (Section 2.0188mm. Ideally.
as this models the behaviour of the material as parts begin to deform plastically.e. a perfectly plastic relationship was assumed.5. the axial force at which the displacement appeared to tend towards inﬁnity (i. 38 . The model also assumed ﬁnite deformations.2 Results Figure 24: Plot of relationship between dimensionless variables The forcedisplacement plots may be found in Appendix B. This is more representative of reality. where ultimate failure was assumed at yield (curve (2) in Figure 23). but more computationally intensive than assuming inﬁnitesimal strains and small displacements (where deformations are considered only relative to the original geometry). This was a conservative assumption to make. The data used be found in Table 8 and the resulting graph from plotting the dimensionless groups is shown in Figure 24. For each analysis. Nonlinear analysis required a stressstrain curve to be assigned to each material to inform Strand how the material would behave beyond yield.used. 5. as all the materials considered undergo plastic work hardening eﬀects that improve capacity before ultimate failure. so to simplify the analysis. Material stress strain relatioships are often complex (see curve (1) in Figure 23). where the model progressively considers the deformed shape from the previous load increment. the ultimate load) was estimated from the graph (using the weaker 6noded joint) and the dimensionless variable values from Equation 23 were calculated. until a limit is reached where deformation will continue indeﬁnitely with no increase in applied force.
006497 S275 14.6.4.0387 0.0 205 0.5 9.0194 0. as its inﬂuence on the relationship was not rigorously tested.5 430 0. The linear polynomial of Figure 24 can therefore be simpliﬁed by ﬁxing the considered radius size as a constant.7 1.1 2.3 2.7 AL6082 205 2. Table 9: Comparison of theoretical and practical joint capacities Material fu (MPa) t(mm) r (mm) Lab Faxial (kN) Theoretical Faxial (kN) AL6082 205 1.0 3.0258 0. the diﬀerence in capacity between hexagonal and pentagonal joint arrays was minimal. = 0. (25) (24) The forcedisplacement curves (Appendix B) indicated that for the theoretical stress analysis.005421 S450 19. The results shown in Table 9. 39 .005929 AL6082 8.5 6. the pentagonal joint capacity was always the greater of the two in each test.0387 0.0258 0.5 and rearranging.13t + 26.005421 S275 15. were compared to to the theoretical equivalents.3 Discussion Figure 24 indicates some form of linear relationship may be established between the two variables.6 6.0044 .005847 S275 16.7 3.0 550 0. calculated using Equation 25.0 1.2 2.5 77. Faxial t 0.5 205 0.0 430 0.0258 0.005441 AL6082 7.5) fu × 77. the tensile failure values for the hexagonal value. Faxial = fu (4.7 1.0 77. It is unreasonable to include the radius in the expression calculated.0 7. taken from Figure 21. Intuitively.0 205 0. while for greater thicknesses it provides a 20% underestimation of their capacity.0 430 0.0194 0.006466 S450 17.0 550 0.5.006466 5. There are a number of reasons that the theoretical estimations become overly conservative. though not comprehensive.2 3.1 As the pentagonal joint results were considered unreliable for the physical tests. in contrast to the practical test ﬁndings discussed in Section 5.Table 8: Nonlinear joint analysis results Material Faxial (kN) t(mm) fu (MPa) t/r Faxial /fu r2 AL6082 6.005847 S450 21.0533 2 (77.5 550 0. indicate that for low joint thicknesses the theory accurately predicts the failure force.43) .0194 0.0387 0.
6 Spreadsheet development This section of the report describes how the FEA based package for analysing geodesic domes works. First. design calculations Start FEA_Input Force vector {F} FEA_Solver Displacement vector {U} FEA_Output End Figure 25: Flow chart of design process 40 . postprocesses the displacement results to give member forces in the structure.1 Package structure The package was broken down into three Excel ﬁles. These ﬁles guide the user through the stages of performing a dome analysis and design. When the joint as a whole is loaded.A possible cause is that isotropic work hardening eﬀects were neglected in the theoretical model (see Figure 23). as well as limitations of the package. “FEA_Output”. A tradeoﬀ between computational eﬃciency and accuracy had to be made. as a single spreadsheet became unwieldy to run due to the large amount of data being processed. with reference made to the theory of Section 3. The spreadsheet subsequently calculates the force vector which is read into the second spreadsheet “FEA_Solver”. and it is possible that the FE mesh was not ﬁne enough to obtain good results. it reaches a state of equilibrium. where the structure stiﬀness matrix is formed. the user deﬁnes a geometry and loading case on the structure in “FEA_Input”. A ﬁnal possible reason for the reserve strength of the joint is that the loaded segment area was conservatively assumed. The worst case force produced is utilised in a series of calculations that allow the user to optimise the structure to carry the load and make savings on materials. Equation 12 is then solved for the structure displacements. As not all the loads on the joint may be of the same magnitude. 6. Additionally. The features that make the layout user friendly and accessible. The ﬁnal spreadsheet. and therefore have a greater capacity of resistance. the boundary conditions assumed in the model represented a worst case. and in reality limited displacements take place at the assumed boundaries of the disc segement modelled. and this contribution became more pronounced as the thickness of the joint increased. are also discussed. materials and forces Axial member forces. Define geometry. it logically follows that larger bolt arrays may be supported by a larger area of the joint than that assumed.
The user can also specify the depth and area of the dome that is loaded. imposing diﬀerent parameters on the calculations. described below.5.3. Additionally. and the various worksheets tabs link to these parameters to calculate forces. loads and partial safety factors. this page shows which portion of the dome is considered loaded.3. A Gaussian elimination method is then used to solve for the displacement vector {U }. • Soil: the nodal forces in the structure due to soil loading are found using Rankine’s theory for earth retaining structures as described in Section 4. • Wind: this worksheet calculates the worst case eﬀect of a wind pressure acting on one face of the structure. 6. Various soil types may be selected from. 6. • Graphs: the level of soil applied to the structure above the base of the foundations is calculated and represented graphically on this worksheet. The methodology of this was previously discussed in Section 4. • Summary: this is the only worksheet in the FEA_Input ﬁle requiring user interaction.1 to form Equation 12.4 for details). materials.1 Stiﬀness matrix assembly The spreadsheet ﬁrst reads in the force vector and speciﬁed parameters from FEA_Input and uses the information about the geometry to scale the coordinates in the “nodal coordinates” worksheet. The main output of the spreadsheet is the force vector.geometry. The force vector is calculated based on the summation of various factored loads. All the variables are deﬁned here. with no user input required.3.3. This in turn is used to calculate the total length L of each member 41 .6. The solver spreadsheet runs automatically when opened. based on a userdeﬁned wind load (see Section 4. • SW & Imposed: the nodal forces due to the self weight of the structure and due to any speciﬁed imposed loading are calculated in this worksheet. The spreadsheet consists of ﬁve worksheets.3 FEA_Solver The FEA_Solver spreadsheet emulates the stiﬀness method discussed in Section 3.3.2 FEA_Input FEA_Input is the spreadsheet in which the user deﬁnes all the parameters of a particular dome .
Y or Z directional DOF in the global coordinate system. referring to whether it was an X. For example. [kglobal] (Eqn 8) [T]T [klocal] (Eqn 6) [T] (Eqns 9 and 11) Figure 27: Assembly of the global stiﬀness matrix Each local element stiﬀness matrix [klocal ] is calculated using Equation 6. Lz (as shown in Figure 26). but the system was chosen to add clarity to which degree of freedom was under scrutiny . In order to assemble the full structure stiﬀness matrix from the 165 global element stiﬀness matrices without resorting to a programming script. The knowledge of how the elements link up was a feature builtin to the spreadsheet. “312” would refer to node 31’s Y directional DOF.Figure 26: Calculation of project member lengths and the axis projected lengths Lx . The remainder of the number (disregarding the last unit) describes which node the degree of freedom belongs to. as described below. based on the element and node numbering scheme used by Strand7. Using the projected member lengths. 23 42 . Ly . The numbering scheme used for degrees of freedom may appear abstract at ﬁrst. [R0 ] and resultantly [T ] are found for each element using Equations 11 and 9 respectively. “2” or “3”. several intermediate stages were used in Excel to get the data in a form that allows the summation of cells with common values. which may be transposed to ﬁnd [T ]T . Equation 8 is used to determine the global element stiﬀness matrix (see Figure 27).the last number is always “1”. Finally. using predetermined information about how elements and nodes link together in the structure23 .
This can be thought of as fully assembled in terms of columns but not rows (Figure 28). Figure 28: Intermediate matrix assembly step 1. An Excel function known as “SUMPRODUCT” was used to achieve this: for each cell in the 183 × 183 structure stiﬀness matrix. (26) 3. This produces a partly assembled version of the structure stiﬀness matrix. For each element. MMULT . Once a [C] matrix was produced for each element.2) and look in the above A Boolean matrix is a matrix with entries from the Boolean domain B = {0. 1965]. the SUMPRODUCT function would seek its reference position (2. whereas the full list of 183 degrees of freedom were included for the columns. It then summed together the cell values for all the matches and outputs the result in the structure stiﬀness matrix.. [C] = [kglobal ][B] .. Figure 29 shows an example of the method for a simple set of matrices. the global stiﬀness matrix and the Boolean matrix were then multiplied together using Excel’s built in matrix multiplication function “MMULT”. This matrix was labelled such that only the rows with DOF associated with the local stiﬀness matrix were considered... 1}. 2. For instance. which shall be henceforth referred to as an intermediate [C] matrix. Such a matrix can be used to represent a binary relation between a pair of ﬁnite sets [Flegg. the calculation seeks the row and column DOF reference of the cell and then looked for each corresponding row and column match in the intermediate matrices. and a one if they do (Figure 28). it was possible to sum all the matching rows in the assembly to complete the full 183 × 183 structure stiﬀness matrix. producing a zero if they do not match. Excel checked each cell’s row and column reference. if the value of the assembled structure stiﬀness matrix circled in red was of interest. 24 43 . A 6 × 183 Boolean matrix24 [B] was created for each element..
matrices and vectors. see Pearson [2008]. The same principles apply for the calculation of each stiﬀness value in the structure stiﬀness matrix. it would ﬁnd two matches. For more information. making it easier for someone wishing to learn the mechanics to follow (see Section 3. 6.2 Solver At this point. This code was adapted from one presented in Billo [2007].Figure 29: Structure stiﬀness matrix assembly using SUMPRODUCT transition matrices for the same reference numbers. circled black. operating as a user deﬁned array function25 in the form: GaussElim(coeﬀ_matrix. The matching cells are summed to get the ﬁnal value which is output in the assembled structure stiﬀness matrix. 25 44 .3. const_vector).2). and const_vector is set as the cell reference area of the force vector. An array function is a special method of calculation employed when working with arrays. both having been ﬁrst reduced by applying the perimeter ring boundary conditions. the spreadsheet has fully deﬁned the force vector {F } and structure stiﬀness matrix [K] in Equation 12. This solver method was chosen as it is popular and well documented. where coeﬀ_matrix is set to equal the cell reference area of the structure stiﬀness matrix. Excel uses a builtin macro written in Visual Basic for Applications (VBA) to perform Gaussian elimination. Rather than applying one formula to one cell. a formula is applied to an array of cells. In order to solve the system. In this case.
An example calculation print out may be found in Appendix E. The local x’ (Figure 8) axis displacements are subtracted from one another. and the transformation matrix (Equation 11) is applied in reverse for each node to get the displacements back in terms of the elements’ local coordinate system. Information about the member lengths and properties are also read into the worksheet. 2007]. to the guidelines of Eurocode 3 [BSI. The displacements are sorted by DOF for each element. by virtue of Hooke’s law. and. so that pdf or paper based printouts could be made for permanent record. The layout was thereby set out in a printer friendly format. where possible. 6.4. rather than applying the full transformation matrix to the full force vector as described in Equation 9.4 FEA_Output The main purpose of the FEA_Output spreadsheet is to determine the internal axial member forces by post processing the results of the structure stiﬀness method. The methods are entirely equivalent. 6. and to allow the user to tailor the fabrication details of the dome so that a safe and economical design is produced.2 Calculations The calculation worksheet was designed to produce a comprehensive document summarising any analysis and design case. the stress can be found using Equation 4.6.4. The strain in each member is calculated using Equation 3. Finally. Note that the spreadsheet uses [R0 ] to transform each node’s 3 degrees of freedom.1 Postprocessing Figure 30: Calculation of member axial forces The ﬁrst step taken by Excel is to import the displacement vector previously calculated in FEA_Solver. The ultimate design load of the structure was set as largest magnitude axial force 45 . giving the change in length of the element due to the applied external forces on the structure. The worksheet was designed. the axial force in each element is found using Equation 2 (see Figure 30). but the chosen approach is slightly more compact and therefore a better method of displaying the information.
calculated in the elements during post processing26 . The spreadsheet makes checks for safety for the Ultimate Limit State (ULS) and Serviceability Limit State (SLS). The following four design checks were considered: • Joint shearing failure (ULS): this check used the empirical relationship developed in Section 5 to estimate joint capacity. Equation 25 is additionally factored by a material constant to represent variable material strengths in steel. • Bolt connections (ULS): the bolts were checked using guidelines laid out for steel bolted connections in EC3, Section 3: Connections made with bolts, rivets or pins. Checks were made for both shear and bearing resistance of the bolts. • Angle section (ULS): this checked the capacity of the equal angle truss elements used by Vigyan Ashram against failure in tension at their smallest net crosssection, i.e. where holes are drilled for the connections, using the guidelines in EC380, Section 3.10.3. Bending moment checks for the section were not required for the reasons discussed in Section 2.3.3. • Deﬂection limit (SLS): this check was introduced to avoid severe deformations of the structure that would impede its appearance or functionality under working loads. The deﬂection limit was set to L/360, which the Eurocodes presented as the most suitable limiting value for a member clad in a brittle material subject to cracking (EC359, Section 7.2). The worksheet also calculates relevant fabrication details, such as the number of joints and diﬀerent member lengths required to assemble the dome and utilises conditional picture formatting to produce subdivision assembly diagrams for the dome (from Landry [2002]) depending on which conﬁguration is selected. It should be noted that while the fabrication and assembly details can be calculated for many diﬀerent dome conﬁgurations, the FE solver method currently only considers the V3 5/8ths dome (see Section 4.4). The expanded version of the fabrication details were created in anticipation that further work may be carried out on the package to expand its analysis capability to any type of dome. At present, the worksheet has warning features built in to inform the user if they attempt a design mismatched from the analysis capabilities.
6.5
Discussion
Before the accuracy of the spreadsheet based analysis is compared to the benchmark results set by Strand7, it is worthwhile discussing some of the features that make the spreadsheet package userfriendly and well suited to the aims of this project.
The reason this assumption could be made without regard to whether it is tensile or compressive was covered in Section 5.2.
26
46
Firstly, the stepbystep layout of the spreadsheet (Fig. 25) leads the user through distinct stages of modelling, analysis and design. By keeping to this format, the layout of the spreadsheets was orientated towards explaining the purpose of each step. While it is entirely possible to perform an analysis using this package without indepth understanding of the inner workings of the calculations, these were left visible and accessible to the users interested in learning how the analysis works. To aid with understanding the FE method, a userguide was produced to accompany the package, explaining the basis of the FE method and how the spreadsheet emulates it. Black box steps were avoided wherever possible in the spreadsheets. For example, the structure stiﬀness matrix may have been more eﬃciently assembled using a VBA macro, but the stepbystep approach implemented in the spreadsheet was felt to be more suitable, as it demonstrates the process with greater clarity. The only exception to this was Gaussian elimination, which had to be performed as a Visual Basic script due to the looping nature of solving a problem using this method. Use was made of Excel based features such as conditional formatting to highlight cells of interest (e.g. passed or failed calculation checks) or to add conditional warnings if the user attempts something inadvisable (e.g. spacing bolts closer than the recommended limit). Additional advice was provided to the user using Excel’s cell commenting feature, where hovering over a commented cell brings up an information box with additional dialogue explaining its function. Table 10: Spreadsheet based FEA results Case Max. Axial Forces (kN) Max. Disp (mm) % Relative error Tensile Compressive ue (cf. Table 3) Case I 4.2329 3.0366 0.8960 (Z) 9.998 × 10−5 Case II 10.2997 69.7237 6.8789 (Z) 2.907 × 10−5 Case III 33.3654 102.4467 7.8202 (X) 3.836 × 10−5 Case IV 1.0610 1.1811 0.2248 (X) 0.000 × 10−5 Case V 2.1607 1.5500 0.4570 (Z) 21.188 × 10−5 Combination I 21.8353 99.5682 9.9605 (Z) 4.016 × 10−5 Combination II 39.6277 143.9114 13.0222 (Y) 0.768 × 10−5 Combination III 7.7372 5.9747 1.5108 (Z) 6.619 × 10−5 The results of the spreadsheet FE method are shown in Table 10. The boundary conditions, forces and design assumptions made were identical to the case described in Section 4.5. The relative error for each load case was found by normalising the maximum Excel displacement ue by the maximum Strand7 displacement us (from Table 3) and expressing it as a percentage: error = 1 − ue × 100 . us (27)
47
For each load case, the relative error between Strand7 and Excel was found to be negligibly small, indicating that the spreadsheet based FE method accurately imitates the commercial package. The computation time to reach a solution was signiﬁcantly longer in Excel (typically 3060s in the spreadsheet cf. to 3s in Strand), as may be expected, but as the point of the exercise was not one of solver eﬃciency, this fact is perhaps unimportant.
7
Conclusions
Returning to the objectives deﬁned in the project plan, it is possible to assess how completely the project objectives have been met. The main goal of this project worked towards producing a Microsoft Excel based structural analysis method, to give Vigyan Ashram the inhouse capability to perform structural analysis. This has been successfully met, with a spreadsheet based system that reproduces the results of an considerably more expensive commercial ﬁnite element package using the structure stiﬀness method. A wide variety of diﬀerent load types and geometries can be applied to the dome, making the package a versatile analysis tool. The system has been designed for ease of use, even for someone with a minimal understanding of the method, while also leaving the mechanics of the method clear for those who wish to learn it. The objective to produce fabrication and design information has also been successfully integrated into the package, leading the user from the stage of forming and analysing a dome model through to using the results of the analysis to design safe joint connections. A linear relationship was found between material properties and estimated failure strength of the dome’s joints. Though this appeared to correlate well, time limitations prevented a thorough testing of the inﬂuence of disc radius as a variable. The physical tension tests performed were good for comparative, empirical studies, and the results conﬁrmed the theoretical relationship developed was a good match. However, the accuracy of the values was questionable as they did not reproduce real conditions, and more work would ideally need to be carried out to conﬁrm the validity of the relationship. The results of the Strand7 analysis were indicative of the problem in the Water Bank project using subterranean domes; very large compression forces became prevalent in the structure, especially in cases of asymmetric loading.
7.1
Further work recommendations
• Expansion of the spreadsheet method to consider more dome conﬁgurations. Currently the analysis focuses on the most common conﬁguration (V3 5/8ths), but with work this could be expanded, oﬀering more ﬂexibility in terms of fabrication and
48
For example. easier to fabricate or stronger. 28 http://www. it may be necessary to investigate the implementation of node ordering algorithms to minimise matrix bandwidth [Cuthill and McKee. so that the software cost associated with the program is eliminated entirely. Alternative joint designs may prove to be cheaper. 1969].the size of structure that can be produced27 . conservative assumptions were made to simplify the loading and the capacity of the structure that may need to be reviewed for this special case.org/ 27 49 . • In depth investigation of the Water Bank project subterranean loading case. the contribution of any passive resistance in the soil was neglected and the testing of the joints focused on tensile rather than compressive failure. Though the analysis and designs tools discussed in this dissertation will produce a safe design for a subterranean dome. • Work on improving the eﬃciency of the spreadsheet solver. • The modiﬁcation of the spreadsheet analysis method to work on a freeware package. this limits the size of dome that can be achieved using a V3 dome frequency. • Investigation into the design of the dome to improve load bearing capacity. As member lengths for lower frequency domes will fail in buckling if particularly large radius domes are planned. For example.openoﬃce. For higher order problems. discs were the only joint considered as they are the design which Vigyan Ashram currently use and are set up to manufacture. such as OpenOﬃce28 .
1965. 1969. H.G Coutie. Billo. Reducing the bandwidth of sparse symmetric matrices.geometer. A practical guide to tensegrity design.A manual for Engineers and Architects. J. URL http://www.in/nicee/IITKGSDMA/W02. Dowrick. John Wiley and Sons. and F. Kumar. 2007.mapsofindia. BSI.ac. 1988. Cuthill and J.com/tenseg/book/revisions. third edition. Fiona Cobb. IS875: Wind loads on buildings and structures. Arlington. URL http://www. Spon Press. 2004. No geodesic structures for quake site.htm. 2007. Structural Analysis.pdf.iitk. Natural hazard map of india.ac. 2003. Coates. Infobase. http://www. Kumar. 2007. Ltd. E. Blackie. Krishna. Blackwell House. In Proc. F. N.tripod. R. [Last Accessed: 28th February 2009]. seventh edition.pdf. 2007.org/mathcircles/geodesic. ACM. P. J. 1987. URL D. [Last Accessed: 1st March 2009]. and K. 1978. Understanding ferrocement construction.M.C.iitk. Virginia. pages 157–172.References N. 2000.html. Earthquake Resistant Design . Wind storms. [Last Accessed: 28th February 2009].K. Bhandari. MFK Group. Hartog. Chapman and Hall. 50 .in/nicee/IITKGSDMA/W03. Newspaper Article. Kong. University of California. October 13th 1993. P. 24th Nat. Gedoesic domes. P. Conf. Graham Flegg. URL http://bobwb.pdf. [Last Accessed: 19th February 2009]. Indian Express. J. Craig’s Soil Mechanics.M. Boolean Algebra and its application. URL http://www. Structural Engineer’s Pocket Book. M. second edition. R. Davis. Krishna. Robert William Burkhardt. damage and guidelines for mitigative measures. E. Bhandari. [Last Accessed: 18th January 2009]. McKee. T. John Wiley and Sons. 1984. Craig.. and K. VITA.com/maps/india/naturalhazard. 2007. Excel for Scientists and Engineers – Numerical Methods. Structural Eurocodes.
Computer methods in applied mechanics and engineering. Mechanics of Fluids. W. The structural making of the Eden domes. Erwin Kreyszig.com/indiamaps/annualtemperature. Space Structures 5. URL http://www. Annual temperature map of india. Space Structures 4. 1979. second edition. 8th March 2009]. 11:2–10. Makowski.com/formula.vigyanashram. Knebel. Building Speciﬁcation October Edition. 1981. Kenner. 1972. 2002. [Last Accessed: 27th February 2009]. SanchezAlvarez.L. 2004. P. 1984. second edition. Introduction to the analysis of braced domes (ed. 1:245–254. pages 387–412. Sarshar. Z. Makowski). Bernard Massey. Z. Meek and S. J. Geodesic math and how to use it. 2006. P. Spon Press. New Zealand.com/. Pakandam and B. Tarczewski. Loading behaviour of some types of tensionstrut domes. Braced domes. 72:57–75. S. Makowski. Analysis. 1989. Analysis. J.mapxl. Zimmermann.html. Space Structures 5. Review of the development of geodesic domes (ed. John Wiley and Sons. pages 86–95. Mullord. and R. 1993. and S. K. URL Z. 2002. Dome formulas. University of Auckland. Inc. Building with Steel. Tara Landry. Morgan.html. Michael Joseph. R. S. 51 . seventh edition. [Last Accessed: 18th January 2009]. 2:1209–1218. MapXL. London. 1984. 2003. University of California Press.desertdomes. Rebielak.D. Design and Construction of Braced Domes. URL http://www. S. [Last Accessed: http://www. Vigyan Ashram: Rural development education system. 1:359–368. Large displacement analysis of space frame structures. 2006. pages 37–46. 1984. Design and Construction of Braced Domes. Kardysz.S. Kalbag. Makowski. Comparison of the behaviour of 3 types of braced dome. J. Design and Construction of Braced Domes. UK. Advanced Engineering Mathematics. Battersea College of Technology. UK. Analysis. S. M. Z. Granada. Steel Space Structures. University of Surrey. 2002. Makowski). Makowski. 1965. Motro. Pitman.. H. Loganathan. The Elements of Structure. 1998. Steel space structures. ninth edition.
URL V. An analysis of braced domes. http://www.V. Level 3 stress analysis. D. 2007. Space Structures 4. Martin’s Press. Palgrave Macmillan. Basic and Applied Soil Mechanics. 1984. 2nd International Space Structures Conference. Engineering Mathematics. Shah. Richter. USA. second edition.S. Jon Trevelyan. Quintas and J. pages 534–549.L. New Age International Publishers. Array formulas. Thomas T. Zung. [Last Accessed: 23rd February 2009].R Rao. St.S. 2000. Taylor and Francis.com/excel/ArrayFormulas. Ranjan and A. New York. The application of excel in teaching ﬁnite element analysis to ﬁnal year engineering students. Introduction to the analysis of braced domes (ed.C. Buckminster Fuller: Anthology for the New Millennium. 2001. Proceedings of the 2005 ASEE/A aeE 4th Global Colloquium on Engineering Education. Analysis. K. 2007. Avila. Space structures development from early concept to temcor geodesic domes. [Course lecture notes]. ﬁfth edition. Soil Mechanics and Geotechnical Engineering. 2001. Design and Construction of Braced Domes. Stroud. A.cpearson. 52 . Horst Friedrich Rolly. pages 175–244. Pearson. Kian Teh and Laurie Morgan. Germany. A.aspx. Shroﬀ and D. K. 1:265–274. M. 2003. Peter Lang. Earthquake Disaster Management. G. Makowski). Frankfurt. 1975. 1993. J. 2005. 2008. L. SanchezAlvarez.
5mm disc 53 .A Project Gantt chart Figure 31: Project Gantt chart B Joint forcedisplacement curves Figure 32: Stress analysis forcedisplacement plot for 1.
Figure 33: Stress analysis forcedisplacement plot for 2.0mm disc Figure 34: Stress analysis forcedisplacement plot for 3.0mm disc 54 .
Appendix C Marek Kubik (Durham Univeristy) Structural Analysis of Geodesic Domes Project.3. India. such as the Pabal dome. The relevant clauses to each calculation are given. clause 5. k1 = 1.2 m/s Wind Loading The following calculations determine the basic windspeed velocity for a geodesic dome structure built in an area nearby to Pabal.0 x 1. and Vz = design wind speed in m/s at height z.1) Combining factors for wind pressure contributed from two or more building surfaces can allow a reduction factor because wind pressures fluctuate greatly and do not occur simultaneously on all building surfaces. such as a cyclone shelter. Assuming a standard design life of 50 years. hence. k3 = topography factor (see 5.0 (Table 19. community building or water tank): k4 = 1.0 (Table 1. Sheet No. the Indian windspeed map gives: Vb = 39.4) Kd = 1.0 x 1.3. otherwise design pressure will become dependant on geometry.2) Basic wind speed is based on peak gust speed averaged over a short time interval of about 3 seconds and corresponds to 10m height above the mean ground level in an open terrain (Category 2). Sheet No. pd = 1.1) Pressure coefficients are a result of averaging the measured pressure values over a given area.0 x 1. Project No.0 x 1. The decrease in pressures due to larger areas may be accounted for with a reduction factor. IS875 (Part 3): 1987.0 x 1. The calculations are guidelines from the Indian Standards. Checked.1) Assuming no hills. Where: Kd = Wind directionality factor Ka = Area averaging factor Kc = Combination factor (see 6.0 x 0. Assume that no reduction factor applies to Pabal dome for simplicty.2). 001 MK Feb 2009 By. k2 = terrain roughness and height factor (see 5. the wind directionality factor is: Kd = 1. p z 0.3. assuming wind action on the Pabal dome acts in only one direction (worst case): Kc = hence. 002 MK Wind calculations for Pabal dome Wind calculations for Pabal dome hence. pressures for buildings less than 10m high may be reduced by 20% for stability and design of the framing.584 N/m2 Design wind speed modifies the basic wind speed to account for terrain effects and the actual height of the building under consideration: V z V b k1 k 2 k 3 k 4 The design wind pressure pd can be obtained as.0 (clause 5.0 (clause 5.8 = 31. and any assumptions are stated throughout. the correlation of measured values decrease and vice‐versa.2. Date.13) For circular or near–circular forms. However.4) The wind pressure at any height above mean ground level is obtained by the following relationship between wind pressure and wind speed: Basic wind speed (clause 5. Design wind pressure (clause 5. However.3.3.0 m/s pz = wind pressure in N/m2 at height z.14) COPYRIGHT © COPYRIGHT © .0 (clause 5. Project No.6V z Where: 2 From figure 1.584 = 0. assuming the height of the structure does not exceed 10m: Vz = 39. Checked. Date. NOTE: The wind speed may be taken as constant upto a height of 10 m. with well scattered obstructions: k2 = 1.3. Feb 2009 By. Appendix C Marek Kubik (Durham Univeristy) Structural Analysis of Geodesic Domes Project.3.4.3.0 x 1. pd kd ka kc p z Where: Vz = design wind speed at any height z in m/s.4. Design wind speed (clause 5. clause 5. cliffs or enscarpments that channel the wind are nearby: k3 = 1.3) k4 = importance factor for the cyclonic region (see 5. Wind Loads for Buildings and Structures. As the area becomes larger.0 (Table 2.0 x 0.1).584 N/m2 1.3) pz = 0.2. k1 = probability factor (risk coefficient) (see 5.3) Assuming the structure is of normal importance (non‐industrial and not of post cyclonic importance.0 (clause 5. clause 6.1) Assume that Pabal dome is built in category 2 open terrain.3.3.4).
.
where possible.6 7. follows Eurocode 3. based on calculations that provide numerical proof of adequacy of the connection details.00 1. the most significant parameters and results of which are summarised below. Appendix 2. Introduction 2. Feb 2009 09001 002 By. Fabrication Data 7.51 7.75 3.Example Calculations 1. Summary of FEA analysis Cladding density Cladding thickness Elem section type Young's Modulus E Elem unit weight Joint weight m N/m 2 22 0. These components include all the truss elements. Summary of FEA Analysis 5. Project. Forces Applied 4. Dome Geometry 3. Date. Checked.7 6. Sheet No. Load Self Weight Soil Wind Imposed PSF 1.22 kN/m3 kN Soil loading indicates whether the dome was analysed under conditions of partial or full burial under some type of soil: Soil type Soil depth Angle buried Black Cotton 3. with all formulae and tables referenced where this is the case.Vigyan Ashram Marek Kubik (Durham University) Pabal Dome Fabrication Details Project. Appendix E .93 kg/m2 2 kg/m Largest compressive force: kN @ beam @ beam 164 131 COPYRIGHT © COPYRIGHT © . The structure was analysed using a Finite Element Analysis package. Checked.2 1 kN/m3 kg per joint kN/m3 The following section provides a summary of some of the key results determined by the FEA solver: Lagest positive displacement: Largest negative displacement: Largest tensile force: 0.78 1.05 0.15 0. as determined by the wind loading code IS875 (Part 3).59 m deg 180 4 Load Combination Factors Load combination factors allow the user to weight the aforementioned loads before they are combined to form a combination load case that is used in the FEA analysis (factors of zero indicate the type of load was ignored).00E+11 77. Dome Geometry Dome type Dome radius Expand base to radius? V3 5/8ths 3 m yes Appendix E . Contents 1. Calculations 6. The connection design. Sheet No.Example Calculations Imposed and Wind Loading MK Wind loading refers to the total force acting on a given face that was considered in the design. Vigyan Ashram Marek Kubik (Durham University) Pabal Dome Fabrication Details Project No. By. Project No. Feb 2009 MK 09001 001 Date.35 0. Forces applied Self Weight The following forces contributed to the self weight of the structure. Imposed loading is an allowance for nonpermanent forces (such as worker access for cleaning or painting). joints and the weight of cladding: 4. Introduction The following document provides fabrication details for the Pabal Dome designed by Vigyan Ashram. Imposed load Wind force Soil Loading 0.05 25x25x3 2.0 mm mm @ DOF 451 @ DOF 13 kN Coarse mesh Fine mesh 1.
5mm. Deflections are not excessive under Servicability Limit State (SLS). Spacing of bolt holes for members Limits: 12 12 22 24 Spacing of bolt holes for disc e1 e2 p1 p2 12. The disc diameter is 77. If this is not possible with the number of bolts chosen. Joints are constructed from isotropic mild steel. 4. Capacity of equal angle channel section used as member. Sheet No. Feb 2009 09001 004 By. By. Members are sufficiently squat not to fail in buckling. so the modulus of the greatest force as determined by the FEA Solver is taken as the ultimate design load: Bolt spacings Appendix E . 5. extreme caution is required altering this radius significantly.5 12.43) Fv . any force that can be withstood in tension will also pass any compression checks).6 10 2 no mm Check that disc joint capacity surpasses the ultimate design load: 14.2 mm kN > 7. Bolts are assumed category A: Bearing type. It is assumed that the critical case for design will always be tensile in the structure (i.5 25 25 mm mm mm mm ≤ e1 ≤ e2 ≤ p1 ≥ ≤ p2 ≥ Joint Capacity 140 140 Disc joint parameters The following joint capacity relationship was developed empirically following physical tension tests and Stress Analysis of the hub joints at the University of Durham: f u (4. Sheet No. and bending moments in the structure are negligible. 4.13t + 26.7 kN OK COPYRIGHT © COPYRIGHT © . Project No. 3. in accordance to EC3 and are designed to withstand predominantly shearing action (due to axial loading). then either a higher grade of bolt class should be chosen or larger connecting members are required. Project. Bolts per connection Controlled bolt size 2 S275 4.05 (a safety factor on the load) Disc thickness t Steel type Bolt class Bolt diameter d0 No. Calculations The following design checks are made in these calculations: 1.Example Calculations 5. Date. Checked. Vigyan Ashram Marek Kubik (Durham University) Pabal Dome Fabrication Details Project No. Fv . Checked. 3. bolt holes and spacings used are homogenous (the same) throughout the structure.Example Calculations MK The array of bolt spacings must conform to the spacing limits shown below on both the disc and the beam. as the joint capacity relationship was designed with this as a fixed parameter.Vigyan Ashram Marek Kubik (Durham University) Pabal Dome Fabrication Details Project. Bolt diameters. Feb 2009 MK 09001 003 Date. Capacity of bolted connection against shear and bearing failure. 2. Rd = γ Where: t= 2 mm fu = 430 N/mm2 γ= 1.7 kN Design Assumptions 1.e. Capacity of dome joint against shearing failure. Ed = 7. Appendix E . 2.
800 Fv . Project.500 2.417 αb = COPYRIGHT © COPYRIGHT © .Rd = α v f ub A γ m2 (EC3 .5 3 x 10 400 430 0.Rd ≥ Fb.Rd then: If Fv.6) (Ultimate load is less than desin bearing resistance .EC3 clauses 3.5 10  1.Rd k α f dt = 1 b u γ m2 (EC3 .7 = d0 2.6 and 3. Rd × n Fv.000 N/mm 2 mm mm2 mm 2 Hence: 0.Rd = 8. Checked.Rd < Fb.8 e2 − 1 . Sheet No. Vigyan Ashram Marek Kubik (Durham University) Pabal Dome Fabrication Details Project No.Rd then: FRd = Fb.5 αd = = αb = MIN = u e1 = 3d 0 12.4) Check that total design resistance of bolts are sufficient: F Rd = 8.800 Where: A= 50 mm2 αv = 0.417 (Assuming that αd is calculated for an end bolt.Rd = 4. Rd Shear resistance capacity per shear plane: F v.Example Calculations Bolt Capacity Two inequalities must be satisfied for safe design of bolts: Fv . Sheet No.7 kN OK Angle Section Capacity Where: d= 10 mm 2 mm t= fu = 430 N/mm2 γm2 = 1. Ed ≤ Fv . which is the worst case) The final check Ultimate Limit State check is that the equal angle sections that make up the truss members will have sufficient resistance to the axial forces in the structure.Vigyan Ashram Marek Kubik (Durham University) Pabal Dome Fabrication Details Project.7) k1 = MIN 2 .6 kN > 7. Feb 2009 MK 09001 005 Date.3 kN/bolt To determine total design resistance of a connection provided by an array of bolts: If Fv.EC3 clause 3. Ed ≤ Fb .930 0.5 f ub = f 1. Project No. The weakest point of these members is through the drilled section.Table 3.4) Hence: k1 = 1. Rd Water Bank Project Calculations MK (Ultimate load is less than design shear resistance . Rd ) × n (where n is the number of bolts per connection) Bearing resistance capacity: F b. Feb 2009 09001 006 By. Appendix E .Table 3. depending on the number of holes.6 fub = 400 N/mm2 γm2 = 1. Date.8 12.7 = 1. By. Checked.4 0.5 (a safety factor from EC367 Section 2) Fb.0 kN/bolt FRd = min (Fv . and several equations exist for determining the section capacity. Chosen section size and other required calculation parameters: Angle section: Steel grade: fu Cross Sectional Area A Net area at bolthole Anet Section thickness t β2 β3 25x25x3 S275 430 157 127 3 0.
F No.0.187383 0. Appendix E . Required numbers No.255167 0. Vigyan Ashram Marek Kubik (Durham University) Pabal Dome Fabrication Details Project No.231598 0. Date. G No. Sheet No. E 25 35 30 30 50 40 30 80 55 30 60 30 30 70 60 30 50 30 60 30 30 60 50 60 30 60 30 90 30 Deflection Limit Check The deflection limit check is a final check to ensure the size of deflections in the structure under the loading conditions assumed will not be excessive enough to lead to cracking or visual distortions.261598 0. Feb 2009 MK 09001 007 Date. Geodesic dome chord factors (unscaled) Dome type V1 V2 V3 3/8ths V3 5/8ths V4 V5 3/8ths V5 5/8ths V6 A 1. Checked.6 kN > kN 7. H No.255167 0.2 kN (3+ bolts) D Member Length E F G H I N u .000 I 0.190477 0.198147 0.31287 0.23179 0. It is not necessary to print this page. 5J 5 6 6 6 6 6 6 6 No.247243 0.162567 B C N u .216628 Geodesic dome fabrication data Dome type V1 V2 V3 3/8ths V3 5/8ths V4 V5 3/8ths V5 5/8ths V6 No. Project.5 1.5 1.000 4 way connection joints 5 way connection joints 6 way connection joints No. I Are all member deflections within the limit (L/360)? 6.25318 0.54653 0.61803 0. 4J 6 20 15 15 20 25 25 30 No.5 430 .000 D 0. Checked.181908 0.261598 0.Example Calculations 7.8 8.Vigyan Ashram Marek Kubik (Durham University) Pabal Dome Fabrication Details Project.230 B 1.Example Calculations 12. Project No. C No.198147 0.2.6 6.247243 0.40355 0. Appendix N u . but the three Ultimate Limit State checks have been passed. Rd = = x 1000 18.40355 0.245086 0. If this check fails.245086 0.29859 0.000 F G 0. Feb 2009 09001 008 By. Rd = 2 . Required 30 55 80 0 0 0 0 0 0 15 6 40 OK (EC3 7.23179 0. By. Rd = β 3 Anet f u γ m2 = x 1000 Check the angle capacity is sufficient: FRd = 14.000 H 0.29524 0.231598 0.1 from Eurocode 3 .41241 0. B No.0 x ( 12.9 kN (1 bolt) MK γ m2 β 2 Anet f u γ m2 1.0 (e 2 − 0 .34862 0. Fabrication data Member Length (m) 1.257 C 0.7 OK 0.063 A 1.5 x 127 x 430 = = = 0.9 fyb (N/mm ) 240 300 480 640 900 2 Equal angle section data fub (N/mm ) 400 500 600 800 1000 2 Angle code 25x25x3 25x25x4 25x25x5 30x30x3 30x30x4 Area (mm ) 157 200 241 187 240 2 Thickness (mm) 3 4 5 3 4 EC331 nominal strength values for hot rolled structural steel Steel Grade S235 S275 fy (N/mm ) 235 275 2 30x30x5 40x40x3 fu (N/mm2) 360 430 291 247 5 3 40x40x4 40x40x5 40x40x6 320 391 460 4 5 6 Assembly diagram of truss elements COPYRIGHT © S355 S450 355 450 510 550 COPYRIGHT © .4 x 127 x 430 = 2.205908 0.245346 0. Sheet No. then the structure is still structurally sound.6 5. A No.1) 30 50 50 60 30 30 130 30 30 65 10 10 60 No.6 kN (2 bolts) 3 x 1000 This appendix tabulates the values that have been looked up for calculations in various parts of the spreadsheet.215354 0.05146 0.225686 0.225686 0.8 10.000 E 0.32492 0.5 x 10 ) x x 14.29453 0. D No.20282 0.bolt class nominal strengths Bolt class 4.5 0.5 d 0 )tf u Appendix E .34862 0.245346 0.198013 0. 6J 0 10 25 40 65 95 120 130 Table 3.41241 0.
This action might not be possible to undo. Are you sure you want to continue?
We've moved you to where you read on your other device.
Get the full title to continue listening from where you left off, or restart the preview.