Escolar Documentos
Profissional Documentos
Cultura Documentos
Final Report
April 7, 2008
Jacob Hammer
CIVE IV
100656672
CARLETON UNIVERSITY
FINAL REPORT
Jacob Hammer Page 2 of 42 14/07/2017
Table of Contents
1.0 INTRODUCTION..............................................................................................................................4
2.0 GOALS................................................................................................................................................4
3.0 BACKGROUND.................................................................................................................................5
3.1 COMPETITION GUIDELINES...............................................................................................................5
3.2 PROPERTIES OF WOOD......................................................................................................................5
3.2 ASSUMPTIONS...................................................................................................................................7
3.3 MATERIAL PROPERTIES.....................................................................................................................9
4.0 MATERIAL TESTS.........................................................................................................................11
4.1 BASE TENSION TESTS......................................................................................................................11
4.2 COMPRESSION TESTS......................................................................................................................14
4.3 SINGLE INTERFACE GLUE TESTS.....................................................................................................14
4.4 DOUBLE INTERFACE GLUE TESTS...................................................................................................17
4.5 MODULUS OF ELASTICITY...............................................................................................................19
4.6 DENTAL FLOSS TESTS.....................................................................................................................20
4.7 TENSION MEMBER CONNECTION JOINT TESTS...............................................................................22
4.8 MEMBER TO MEMBER JOINT TESTS................................................................................................24
5.0 BRIDGE DESIGN............................................................................................................................25
5.1 OVERALL DESIGN...........................................................................................................................25
5.2 LAYER DESIGN................................................................................................................................25
5.3 STRUCTURAL ANALYSIS..................................................................................................................26
5.4 COMPRESSION MEMBER DESIGN....................................................................................................27
5.5 TENSION MEMBER DESIGN.............................................................................................................28
5.6 DECK DESIGN.................................................................................................................................29
5.7 STRUT DESIGN................................................................................................................................29
5.8 PIER DESIGN...................................................................................................................................29
5.9 DESIGN DIFFERENCES IN FINAL MODEL.........................................................................................30
6.0 COMPETITION RESULTS............................................................................................................31
6.1 ANTICIPATED METHOD OF FAILURE................................................................................................31
6.2 METHOD OF FAILURE......................................................................................................................31
6.3 EXPLANATION FOR FAILURE...........................................................................................................32
6.4 IMPROVEMENTS / NEXT STEPS........................................................................................................32
7.0 CONCLUSION.................................................................................................................................33
7.1 PROJECT RESULTS...........................................................................................................................33
7.2 FINAL THOUGHTS............................................................................................................................34
Jacob Hammer Page 3 of 42 14/07/2017
APPENDIX...................................................................................................................................................35
APPENDIX 1 RAW DATA FOR TEST 1........................................................................................................36
APPENDIX 2 RAW DATA FOR TEST 2........................................................................................................37
APPENDIX 3 RAW DATA FOR TEST 3........................................................................................................38
APPENDIX 4 RAW DATA FOR TEST 4........................................................................................................39
APPENDIX 5 RAW DATA FOR TEST 5........................................................................................................40
APPENDIX 6 RAW DATA FOR TEST 6........................................................................................................41
APPENDIX 7 RAW DATA FOR TEST 7........................................................................................................42
APPENDIX 8 RAW DATA FOR TEST 8........................................................................................................43
APPENDIX 9 RAW DATA FOR TEST 9........................................................................................................44
APPENDIX 10 RAW DATA FOR TEST 10....................................................................................................45
APPENDIX 11 RAW DATA FOR TEST 11....................................................................................................46
APPENDIX 12 RAW DATA FOR TEST 12....................................................................................................47
APPENDIX 13 RAW DATA FOR TEST 13....................................................................................................48
APPENDIX 14 READOUTS FROM SAP2000 MODEL..................................................................................49
APPENDIX 14 READOUTS FROM SAP2000 MODEL (CONTINUED)...........................................................50
APPENDIX 15 IMAGES FROM SAP2000 MODEL.......................................................................................51
APPENDIX 16 DESIGN GUIDE...................................................................................................................52
Table of Figures
3.0 BACKGROUND
3.1 ACCEPTED MATERIAL PROPERTIES OF BIRCH...................................................................................9
4.0 MATERIAL TESTS
4.1 SINGLE STICK TENSION TEST SETUP..............................................................................................11
4.2 TRIPLE STICK DRY INTERFACE TENSION TEST SETUP....................................................................12
4.3 TRIPLE STICK GLUED INTERFACE TENSION TEST SETUP................................................................13
4.4 COMPRESSION TEST SETUP.............................................................................................................14
4.5 SINGLE INTERFACE GLUE TEST SETUP...........................................................................................15
4.6 DOUBLE INTERFACE GLUE TEST SETUP..........................................................................................17
4.7 DENTAL FLOSS TEST SETUP FIRST CONFIGURATION......................................................................21
4.8 DENTAL FLOSS TEST SETUP SECOND CONFIGURATION..................................................................21
4.9 SCARF JOINT TEST SETUP...............................................................................................................22
4.10 HALF LAP SPLICE JOINT SETUP......................................................................................................23
4.11 DOWEL PIN CONNECTION SETUP....................................................................................................24
5.0 BRIDGE DESIGN
5.1 OVERALL LAYOUT OF BRIDGE DESIGN..........................................................................................25
5.2 VENEER DESIGN CONFIGURATION..................................................................................................25
5.3 ANGLE REFERENCE CHART.............................................................................................................26
5.4 CORNER JOINT ECCENTRICITY........................................................................................................26
5.5 COMPRESSION MEMBER CROSS SECTION.......................................................................................27
5.6 VENEER REQUIREMENT REFERENCE CHART..................................................................................28
5.7 TENSION MEMBER TOP VIEW CROSS-SECTION..............................................................................28
5.8 DECK DESIGN CROSS-SECTION.......................................................................................................29
5.9 STRUT DESIGN CROSS-SECTION......................................................................................................29
5.10 FINAL MODEL DESIGN....................................................................................................................30
6.0 COMPETITION RESULTS
6.1 JOINT SEPARATION FAILURE............................................................................................................32
Jacob Hammer Page 4 of 42 14/07/2017
1.0 INTRODUCTION
This report is intended to outline to the department the findings of my fourth year
project. This includes the research which has been performed, the resultant
findings, and the analysis thereof.
2.0 GOALS
First, the main goal is to research the materials permitted for construction of a
bridge, and design a bridge to be entered in the Concordia Bridge Building
Competition. The project aimed to provide an analytical approach to the design
and construction of the bridge and document the research that went into it.
Second, the supporting goal was to build a model bridge built out of designated
building materials (see section 3.1) that is capable of supporting 3 kN. After
initial research, this was increased to 5 kN, or just over 1100 pounds force. In this
context, the bridges capacity is the maximum point loading being applied to the
bridge at its centre at the point of collapse, as indicated by the testing machine.
This value was picked as I believe it represents a quantitative threshold for
acceptable bridge performance in the context of a fourth year engineering project.
As well, this value seems to be a threshold for the top 10 positions in previous
years at the Concordia Bridge Building Competition. I decided to aim for an
ultimate capacity goal instead of an efficiency goal because not only is the
capacity worth more in the competition scoring, but realistically, a hefty yet
inefficient bridge is infinitely more functional than a highly efficient bridge with a
low capacity.
3.0 BACKGROUND
While a complete set of rules may be obtained from the CSCE Concordia website,
below is an abridged set of guidelines that directly dictate restrictions on the
bridge:
These are descriptions of the basic properties of wood. They are stated here to
give a baseline for all further remarks and assertions.
Jacob Hammer Page 6 of 42 14/07/2017
Wood has no natural strength ultimate limit when compressed perpendicular to the
grain. After the proportional limit, it will continue to deform and warp beyond
recovery, severely affecting the woods properties in all other orientations. The
conclusion drawn from this is that wood should not be loaded under perpendicular
compression unless necessary.
Parallel to the grain, wood has a fairly reasonable strength. The design values for
wood are typically 75% of the maximum crushing strength for hardwoods and
80% for softwoods. Beyond these design values, the stress-strain curve exceeds
its proportional limit and the wood will begin deforming beyond recovery.
Because the goal of the project is to test the absolute capacity of a bridge, no scale
factor will be used and the maximum crushing strength will be accepted as the
parallel compressive strength for design and analysis. At failure, the fibers will
buckle and snap much like slender columns. The failure will most likely occur at
the members joints. In the few compression tests I attempted in the Instron
machine, the samples broke right at the tips of the grips.
In lieu of published values for tensile strength of wood parallel to the flow of
grain, the modulus of rupture can be substituted. While this value is not a true
reflection of the woods tensile capacity, it will serve as a conservative estimate
viable for design. At failure, two basic models can occur. One is that the fibers in
the wood will elongate and snap (called a tear out). Alternatively, the grain
orientation will permit the member to fail in parallel shear as detailed below.
Perpendicular tensile failures are similar to a parallel shear mode detailed below.
This means that the perpendicular tensile capacity of a wood is a measurement of
the cohesive strength of the lignin. As observed in the published strength values,
its similar to the value of compressive perpendicular loading and as such, it
should be avoided in favour of a more accommodating loading orientation.
This is the basis for the failure mode of tensile members to shear out. This
means that if the grain is not completely parallel to the cut of the member, the
lignin will give away and rather than have the fibers themselves yield and snap,
the surface between the fibers will separate creating a cleavage surface. In
parallel grain tensile members, this surface will shear apart.
Jacob Hammer Page 7 of 42 14/07/2017
As in all materials, the modulus of elasticity is taken as the slope of the stress-
strain curve prior to the proportional limit. Its a measurement of the tendency of
a material to behave elastically before it behaves plastically.
3.2 Assumptions
3.2.1 Shrinkage
Wood will naturally shrink as its moisture content decreases. For this project, the
shrinkage properties of wood will be neglected. This is for two reasons:
Primarily because the sticks have already been dried prior to packaging (although
the box does not indicate to what extend), but also because the shrinkage effect is
proportional to the dimensions of lumber. An educated guess of the sticks dried
moisture content can be made based of its EMC, however this would not affect
shrinkage. The dimensions of the sticks are so small relative to construction
lumber that shrinkage and growth effects are negligible when weighed against any
inconsistencies as a result of the construction process.
It is also assumed that during the testing, construction, and transportation (to the
competition) process of any test samples or bridge components, there will be no
amount of precipitation or condensation present in enough quantities to be
noteworthy. While contact with glue will be a major part of the construction
process, it is assumed that any local expansion caused by the introduction of
moisture will merely help the joint fit together better, if it has any effect at all.
3.2.2 Warping
In testing the bridge, it is loaded with a constant point load over the span of a few
minutes until failure. While this would result in a short term loading condition,
the structure is being loaded to failure and as such, it not being designed to
accommodate an extended life. Because of this, the K D factor for this bridge is
1.0.
It is assumed that the bridge is designed and will operate under dry service
conditions. Because of this, the KS factor for this bridge is 1.0.
As advertised on the companys website1, the wood used for these sticks has no
treatment of any kind. Because of this, the KT factor for this bridge is 1.0.
From table 5.4.42, the system factor for tension and compression parallel to grain
in a built up beam is 1.00, therefore the KH factor for this bridge is 1.0.
While the frame has no bearing point lateral stability, the piers provide a wide
stance. This concept was tested in previous years bridges and the result was
found to be that it had a significant impact on lateral stability of the bridge, such
that there was a negligible chance of the bridge failing to either side. Because of
this, the KL factor for this bridge is 1.0.
1
http://www.loew-cornell.com/education/tipsandtechniques/flashpaper/40/index.html, January 6th, 2008
2
CSA-086-01, 2005 edition, page 30
Jacob Hammer Page 9 of 42 14/07/2017
ForsterTM, the company that produces the sticks I am using, has claimed on their
website that all craft sticks they produce are made from 100% birch wood which
has not been treated in the production process. While it is used for furniture and
small fabrications, birch is not a standard construction wood, and as such stock
values for it were not readily available. Research online provided the values
below from the American Hardwood Information Centre3, which are detailed
below in figure 3.1
(all units converted from Psi) Moisture Content
Green 12%
Modulus of Rupture (MPa) 44.1 116.5
Modulus of Elasticity (MPa) 8066.9 14961.6
Compression Parallel to Grain (MPa) 16.3 58.9
Compression Perp. to Grain (MPa) 1.9 7.4
Shear Parallel to Grain (MPa) 5.8 15.4
Tension Perp. To Grain (MPa) --- 6.6
Figure 3.1
a) Results of tests on small clear specimens in the green and air-dried conditions.
Definition of properties: compression parallel to grain is also called maximum crushing
strength; compression perpendicular to grain is fiber stress at proportional limit; shear is
maximum shearing strength; tension is maximum tensile strength; and side hardness is
hardness measured when load is perpendicular to grain.
b) Modulus of elasticity measured from a simply supported, center-loaded beam, on span
depth ratio of 14/1. To correct for shear deflection, modulus can be increased by 10%.
After testing of the axial properties of the popsicle sticks in various configurations
(detailed in section 4.0), I was able to obtain approximate strength values for the
popsicle sticks. Unfortunately, there is no published record of the moisture
content of the packaged sticks. I can assume, however, that the properties vary on
a linear proportionality. Based on this assumption, I can use the values I have
calculated to approximate where the sticks lie on the moisture curve and estimate
the material properties that I was unable to derive directly through testing.
Between green and 12% conditions, the ratio of tensile strength parallel to the
grain to the shear strength parallel to the grain differs by only 0.97%. From this I
can gather that the woods resistance to tensile failure in both fiber snap and shear
out modes essentially increases linearly as it dries. Knowing the tensile strength
of the wood from tests, I can estimate the shear strength within a reasonable area.
3
Wood Handbook, Wood as an Engineering Material, USDA Forest Service
Jacob Hammer Page 10 of 42 14/07/2017
Between green and 12% conditions, the ratio of the compressive strength parallel
to the grain and perpendicular to the grain differ by about 9.5%. While this value
is higher than I would like, it still indicated a reasonable tight correlation. From
this I can gather that the woods compressive strength increases at a steady rate as
it dries, but the perpendicular strength gradually gets stronger in relation to the
parallel strength. To err on the safe side, I will assume the same ratio in green
conditions all the way through resulting in a slightly lower (but not drastically so)
design value at lower moisture contents.
Between green and 12% conditions, the ratio of the compressive strength parallel
to the grain and tensile strength parallel to the grain differ by about 27%. This is
a very high number and can partially be accounted for due to the ratios being
relatively small (2.7 and 2.0 respectively). This indicates that as the wood dries,
the strength gain in compression slows relative to the strength gain in tension. I
also know that as the wood goes from green to 12% conditions, the ratio of the
parallel compressive strength to parallel shear strength differs by 35.7%
(increasing from 2.8 to 3.8). Using these two relationship curves, and the tight
correlation between tension and shear, I can approximate compressive strength by
assuming that (a) the ratio change is linear through the drying process and (b) the
average of the two estimates reflects the most accurate compressive strength.
At the 12% moisture level, I can see that the perpendicular tensile capacity is the
weakest loading configuration and has a practically non-existent capacity in the
green condition. From this I can assume that it is never a good idea to place birch
wood in such an orientation and that I should ensure no sticks are exposed to
perpendicular tensile loading as they would guarantee to be the failure point in the
member. This also means that sticks should be examined to ensure that the cut of
the stick aligns with the orientation of the grain. Some sticks are cut across the
grain since they are marketed as craft sticks and not design to be exposed to such
loading. These sticks must be discarded. In the final design, all sticks used will
be inspected and any sticks with the grain orientation exceeding 10 degrees
deviation will be discarded. Sticks with grain orientation almost parallel with be
used for sections under tensile loading and the rest will be used for sections under
compressive loading.
All that said, the tests indicated that the tensile capacity was well below the
published strength values for green conditions, having a mean tensile stress of
25.7 MPa at failure averaged between the three raw tension tests. This indicates
that the sticks were on average the same strength, and that the testing machine
was accurate in its readings. The difference in the 5% strengths was due to the
way each sample type carried the loading. The single sticks, being only a single
stick with no glue or load sharing, reflect the true strength of the sticks. The built
up members experienced load sharing (the full length glue contributing slightly
more) and thus had higher 5% strength values. While the built up members are a
better representation of the conditions in the bridge members, taking the tensile
capacity from Test 1 gives the most conservative result.
Jacob Hammer Page 11 of 42 14/07/2017
While it was impossible to test the compressive capacity of the sticks, the closest
calculation would be to estimate using the ratios discussed above. If I take the
ratio between tensile and compressive parallel capacity as 3.0, a conservatively
scaled value of its green condition value, then I can estimate the compressive
strength of the sticks as being roughly 4.085 MPa.
While lower than the published values, accepted strength values for the sticks
were found from material tests to be 12.3 MPa tensile and 4.1 MPa compressive.
The modulus of elasticity was found to be 1700 MPa.
Below is a discussion of all testing done so far, and still to be done. It includes
diagrams, observed results (if completed), failure mode discussion, and
observations on how the samples behaved differently from expected. For each
subsection, the test number shown in parenthesis indicates the corresponding test
results in the appendices.
For all tests, results are found in the appendix with the corresponding test number.
This run had three different sets of samples: A single stick which had its centre
narrowed, a set of three sticks with the tips glued, but the centres dry, and a set of
three sticks with their entire lengths glued. The three sets were all run with the
intention of finding the tensile capacity of the popsicle sticks.
Figure 4.1
4.1.1 Single Stick Test [Test 1]
My initial theory was that these samples would fail by tear out at
or very close to the vertex of the necking. Because each side of the
necking was done separately, they didnt completely line up and
the two vertices were usually off by about 1 mm, although I didnt
expect this to have much effect.
Jacob Hammer Page 12 of 42 14/07/2017
The most common failure method was a tear-out. The shear line would start at
the vertex of the necking and would follow the grain to the edge or grips,
whichever came first, and separate. When it reached the grips first, the remainder
of the distance from the shear line to the edge would tear out along the grip in the
direction of shortest distance to the edge. This shear out is due to the alignment of
the wood grain relative to the cut of the sticks. In sticks with perfectly aligned
grain (about 15%), the sample would snap at the vertex of the neck.
Figure 4.1 shows the sample setup and loading configuration. Capacity was
calculated as P / A where A was the cross-sectional area at the narrowest point.
In the end, this batch of test samples behaved almost exactly like
I anticipated. Under the loading, the samples had a notably
higher 5% strength, but the mean capacity was very close to Test 1s. The
observed failure mode was typically having all three sticks experience stress until
one or two of them gave and snapped. The remainder would usually hold up the
stress for a second or two longer before snapping. This shows that for a built up
section, when one component broke, its share of the load would be carried over to
the remaining sticks, but it would break soon-after. Therefore, for any bridge
members, this test demonstrates that I can estimate with fair certainty that there is
a linear relationship between area and capacity. My hypothesis about the effects
of the jaw crushing the sample were also proven, as most of the samples fractured
in the centre without having the failure lines approach or enter the jaws.
Jacob Hammer Page 13 of 42 14/07/2017
Figure 4.2 shows the sample setup and loading configuration. Capacity was
calculated as P / A where A was the cross-sectional area of three sticks. This was
divided by three to get the approximate capacity per stick.
After seeing how the first two tests ended, I assumed that this test
would proceed in a similar fashion. The member would have a
capacity of roughly three times test 1 and break in a shear out
mode.
In the end, this test behaved almost identically to test 2, with one
key difference. Because the sticks were glued together,
compressed, and cured, they behaved as a single member, and
shear planes would traverse all three sticks wherever possible at
failure. Failure would cause all three sticks to break
simultaneously. This shows that all the glue managed to do was
hold the member together, and that the ultimate capacity was still
dependant on the sticks, not the glue.
Figure 4.3 shows the sample setup and loading configuration. Capacity was
calculated as P / A where A was measured as the average cross sectional area of
the members at the centre. This was divided by three to get the approximate
capacity per stick. Because the glue would increase the cross-sectional area by a
marginal amount, I decided to take an average of 10 samples to see how much
larger the cross sectional area was as a result of the glue. I found that the different
was a negligible 4 mm2. This was likely due simply to measurement inaccuracy.
4.1.4 Conclusions
These tests were all designed to gauge the tensile capacity of the popsicle sticks.
There were also designed to approximate how the capacities of build up sections
would be estimated. The difference in 5% strength values between the tests was
notable, however there is a good reason. Test 1 had a great deal of variability in
Jacob Hammer Page 14 of 42 14/07/2017
sample construction (resulting in a larger range of results) while tests 2 and 3 had
less (resulting in smaller ranges), so this would change the 5% ratings (which are
essentially a function of test range). While the 5% ratings werent as close, the
means, a better measure of the relationship of the tests, were very close. This lets
me conclude that for any built up sections, I can approximate the equivalent
tensile capacity as the sum of the sticks involved in it. Tests 2 and 3 showed that
the glue interfaces really had no contribution towards tensile capacity, supporting
my conclusion, and really only works to guide the failure lines between sticks.
This conclusion can only hold provided that the sticks used have no knots or other
critical flaws, and are all of uniform dimension.
Taking the 5% strengths of the weakest test is the safest bet for a design value.
Therefore, I conclude that assuming a tensile capacity of 12.3 MPa will ensure
that my design will almost certainly be stronger than anticipated.
This run had three sets of samples with glued interfaces of different lengths. The
three sets were all run with the intention of finding the shear strength of the glue,
as well as the threshold at where the glue becomes the limiting factor.
Jacob Hammer Page 15 of 42 14/07/2017
My initial theory was for samples with such a large bond area (almost half the
length of the sticks), the limiting factor would be the sticks. I anticipated the
glued portion, having been properly prepared and dried, to act as a single member
and have the sticks break outside the ends of the interface.
For this batch, my theory was mostly confirmed. The sticks broke at the edge of
the joint, leaving the interface intact. While a few of the sticks failed by interface
separation, they were a clear minority. This shows that for a 6 cm interface
length, the sticks are definitely the limiting factor. It also shows that for a built up
member that consists of both dry and glued sections, the failure will almost
certainly occur at the edge of the glue joint.
Following the observations from test 4, my hypothesis for this set of samples was
that they would fail in a similar manner, at the joints, however as the interface
length was shorter, there was a better chance of the glue joint being the limiting
factor and thus the samples failing from interface failure.
This batch followed my theory with the number of interface failures rising from
less than 5% to about 30%. This shows that while the majority of the samples
still failed by the sticks breaking at the joint, the glue interface has reached a point
where it can be the limiting factor if the construction conditions of the interface
are inferior.
Test 5 has an identical configuration to test 4, and as such, section 4.3.1 describes
the process by which the stress capacity for these samples was calculated.
Following the observations from test 4 and 5, my hypothesis for this set of
samples was that while some would fail at the joints, the interface size will have
reached or passed the threshold and that it will now be almost exclusively the
limiting factor.
For this batch, my theory was confirmed. Interface separation failures made up
about 90% of the failure modes, with the sticks themselves showing little to no
signs of elongation, or fracture. This clearly shows that with a 2cm glue joint
interface, the glue has become the limiting factor.
Test 6 has an identical configuration to test 4, and as such, section 4.3.1 describes
the process by which the stress capacity for these samples was calculated.
4.3.4 Conclusions
These tests were meant to gauge the shear strength of the glue, as well as discover
the threshold for when an edge-grain glue joint is weaker than the wood its
connecting and will give first. I discovered that even that for a 570 mm 2 joint, the
glue has a slight chance of failing first. As the joint area decreased, the glue had a
higher chance of failing, and under 200 mm 2 the joint was essentially guaranteed
to fail first. While it would be impossible for all sticks to be connected with joints
larger than 6 cm in the actual bridge (as the sticks are just under 12 cm long),
these tests only observed single shear plane failure. I shall make a tentative
conclusion for now that with a built up member and sticks glued on both sides, the
glue will not fail first. I will test this hypothesis in the next run of tests.
The glue used, made by Mastercraft, advertises a shear strength of over 19.3
MPa (taken from the back of the bottle). My tests showed that this glue
performs well under this advertised capacity. I am unsure as to the conditions
under which this value was decided upon, but since it was provided by the
manufacturer, Im sure that this value was unrealistic and only attained once
under absolutely ideal lab conditions. Sample variance is a product of amount of
glue used, compression applied during the drying process (which can thin the
layer of glue between sticks) and accuracy of the interface dimensions (some
sticks slid up to 2 mm under the compression in the drying process). Because of
this, and the wide range of ultimate capacities under which the interface joints
failed, I cannot conclude for certain what the shear capacity of the glue is. What I
can ascertain is how large I need to make my interfaces to avoid having the glue
fail. From these tests, my conclusion is that an area of over 600 mm 2 (cumulative
between both sides) would prevent a glue separation from controlling.
My initial theory was that for such a large interface area, there would be no
interface failures, and all samples would fail at the edge of the joints.
These samples primarily followed my hypothesis, but there were a few interface
failures. These failures were few and were not clean interface failures observed in
the smaller joint tests, but the wood around the interface failed along with the
glue. I believe this was merely a result of weak wood in the side sticks and the
joints failing as close to the joint as possible. While I assumed for these tests that
two parallel 4 cm interfaces would be the same as a single 8 cm interface, the
truth is that theyre still only 4 cm long, which had several interface failures in the
single interface tests. An improper or unusually large axial force could cause one
of the sides to give and with only half the support, the other side of the joint easily
snaps as joint instantly has to take twice its intended loading.
Figure 4.6 shows the sample setup and loading configuration. The sticks
capacity was calculated as P / A where A was the cross-sectional area of a single
stick.
These samples were assembled as detailed in figure 4.6. Tensile force was
applied to the central stick until failure. The interface at each end was 3 cm long
on either side of the central stick for a total interface area of 570 mm2 at each end.
From the single interface test observations, my theory was that these would
follow the same failure pattern as the single interface tests, with test 8 having a
few more interface failures than test 7, but a fraction of the number in test 9.
Test 8 followed the pattern I was expecting with the single interface run of tests.
Figure 4.6 shows the sample setup. The sticks capacity was calculated as P / A
where A was the cross-sectional area of a single stick.
These samples were assembled as detailed in figure 4.6. Tensile force was
applied to the central stick until failure. The interface at each end was 2 cm long
on either side of the central stick for a total interface area of 380 mm2 at each end.
Similar to test 6, I believed that these samples would have the highest rate of
interface failures for all three tests in this run. At least half of these samples
would fail by interface separation.
Jacob Hammer Page 19 of 42 14/07/2017
Figure 4.6 shows the sample setup. The sticks capacity was calculated as P / A
where A was the cross-sectional area of a single stick.
4.4.4 Conclusions
These tests were meant to verify my conclusions from the single interface test run.
After observing these tests, I can conclude firmly that I was correct. Test 7 had
the same number of interface failures as test 4, showing that while both joints
were sturdy, they could still fail. As the interface area got smaller, the number of
samples that failed by interface separation increased proportionately, confirming
that any interfaces smaller than those in test 4 or 7 would simply be inadequate.
The absolute lowest shear strength found in a sample from Tests 4 to 9 in which
the glue interface failed was 1.65 MPa (found in Test 7). This means that a single
9.5 mm wide popsicle stick glued along its entire length (assumed to be 10 cm) on
both sides would require an axial force greater than 3.1 kN to separate the
interface.
In construction of final bridge members, the largest bond I can form between
sticks is to cut all sticks to the same length (by squaring off the ends) and have
them overlap with half their length (about 45 mm) on either side. Following the
trend of these tests, there shouldnt be any interface failures. In the event the
interface between any two given sticks fails, those sticks would still be attached
on their other side, allowing the member to continue supporting the same axial
load.
This test is designed to test the modulus of elasticity of the popsicle sticks.
A single stick is placed in the grips of the testing machine and loaded at a much
slower speed than the rest of the tests (approximately 1.5 mm per minute).
Readings are taken at short arbitrary intervals for several samples until a
reasonable pattern emerges. These sticks were not narrowed at the centre because
I needed them to last longer before they break, as well as because the test was
more concerned with their behaviour before rupture instead of at rupture.
Jacob Hammer Page 20 of 42 14/07/2017
I expected the sticks to fail much in the fashion of those in Test 1. I also expected
the modulus curve, when graphed, to have a fairly straight correlation ending in
abrupt failure given the brittle nature of wood, as opposed to the curve seen in
metals.
As observed in the chart in appendix 10, the samples had a quickly emerging
pattern. The stress-strain curve of the sticks was a straight line (albeit with some
wiggle due to inaccuracies in recording) that was easily evaluated.
The modulus of elasticity was taken as the slope of the line of best fit, which is
parallel to samples 3 and 4. It is evaluated as rise over run.
4.5.2 Conclusion
Using sample 4, which yielded nearly ideal results, the slope can be calculated
using the equation [ stress ] / [ ( deformation) / length ] = MoE
Assuming deformation along the entire length of the stick (113 mm):
[110.6 - 9.5] / [(3786-243)/113)] = 3224 MPa
Assuming deformation along only the exposed length of the stick (60 mm):
[110.6 - 9.5] / [(3786-243)/60)] = 1712 MPa
In this test, the jaws have interfered with the proper measurement of the modulus.
While the entire length did not deform uniformly due to the effect of the jaws, it is
also incorrect to assume that only exposed section deformed as well. As there is
no way of knowing what percentage of the overall deformations occurred in each
section, the most conservative estimate would be to assume 100% in the exposed
region.
This test is designed to test the tensile capacity of dental floss and its feasibility as
a cable tie to be used in a popsicle stick bridge.
A braid of 10 strands will be weaved into a simple braid and will be coated in
glue. Tests 2 and 3 have already demonstrated that the white glue will have
almost no contribution in tensile capacity and thus will only serve as cohesion on
the braid for testing.
Jacob Hammer Page 21 of 42 14/07/2017
To test this, two anchors of 3 sticks glued together and dried will
have a hole drilled through them and a braid looped between them.
The anchors will be placed in the grips of the Instron and tested in
tension. The machines readings will be used to determine the
maximum tensile capacity at failure should the braid snap. Figure 4.7
As predicted, testing resulted in local tear-out failure. Because the anchors had
grain parallel to the line of action, the braid easily tore out a path between the
fibers. While further testing of the dental floss is required, this experiment has
made it apparent that using dental floss as a cable when its line of action is
parallel to the grain of the wood is an entirely bad idea.
Figure 4.8
A second experiment was performed to accommodate for the grain
orientation as illustrated in Figure 4.8. The floss rope was tied around
two smaller blocks with their grain perpendicular to the floss. The trial
sample of this test gave very unfavourable results, with the 5-braid
strand taking 280 N with an accompanying deformation of 11.9 mm
before the braid snapped. While this is an impressively large loading for
such a small area, this means that using the dental floss would be overall
unproductive on the bridge as it would deform too much to be of any use
in small braids. A braid sufficiently large would require too large a hole
drilled in the bridge compromising the integrity of the member.
enough to control tensile capacity, its size would significantly impact the
geometry of the sticks. In compression, the floss could contribute nothing and the
sticks already low capacity of 4 MPa needs as much cross-sectional area as it can
get, with the floss only reducing it.
Because of the member size restrictions of the bridge, the tension member must
be made of two smaller sections that are jointed together prior to testing. In order
to do this, the two base members must have a prefabricated interface that can be
easily assembled prior to construction. These tests explore the possibilities
considered for construction this joint. It should be noted for all tests below that
glue yields strongest bonds when used between edge-grain surfaces (two sticks
side to side), and weak bonds when used between end-grain surfaces (two sticks
end to end).
All samples failed by interface separation leaving the two halves fully intact, With
interface areas well below my calculated requirements above. The samples had a
superior strength and showed that the glue used is very strong in perpendicular
separation, making this a very feasible candidate for the tension joint. My
hypothesis was actually opposite of the results, with the shallower angles yielding
worse results.
Jacob Hammer Page 23 of 42 14/07/2017
Figure 4.9 shows the sample setup and loading configuration. The joints capacity
was calculated as P / A where A was the cross-sectional area of the glue interface,
calculated as [w * h / sin ()].
Similar to the straight interface test this interface would use the
edge-grain strength property of glue to its advantage. The tension
member would be made of three Z-shaped pieces, with the inner
piece being simply to extend the length and the outer two pieces
having the joints for them to connect to the two compression
members. This setup would require three pieces instead of two
because having two members connect at a lap joint wouldnt do
much to reduce the sub-members lengths. Figure 4.10 shows how
these pieces would be assembled.
These samples were essentially much wider versions of the coupon samples used
in tests 4 to 6, and yielded very similar results. The most significant similarity
was deformation. While these samples had somewhat lower capacities, the
deformation result was significantly lower, making this a prime candidate for a
rigid joint.
Figure 4.10 shows the sample setup and loading configuration. The joints
capacity was calculated as P / A where A was the cross-sectional area of the glue
interface, calculated as [w * L].
4.7.3 Conclusion
While the Scarf Joint test showed significantly better capacities, the Lap Splice
had a fraction of the deformation. I can build the tension member to have much
larger areas for the lap splice joint, permitting it to reach higher capacities, but the
results from both tests clearly indicated that the scarf joint had a 34% increase in
capacity, but cost a 406% increase in deformation, something I cannot afford to
have in the tension member of my bridge. Therefore, I shall build the tension
member joint out of a lap slice, using the calculate 5% capacity of 1.4 MPa.
This test is designed to find the best method for attaching the
individual members to each other to construct the final bridge
design.
This joint was a complete failure. While the dowel has been perfectly rounded
down and the assembly fabricated as per figure 4.11, the two outer connections
merely slid apart as the assembly was pulled. It ultimately failed when the length
L (the distance from the tip to the hole, cut 1.5 times the pin diameter from the
tip) tore out. This joint has proven to be a poor choice without proper fastening,
which would be infeasible given materials provided. Furthermore the pin joint
would be required to be used when connecting two members near the tip, making
it infeasible and prone to tear-out failure.
4.7.3 Conclusion
A pin joint will not be used on this bridge, and instead members will be connected
by an overlapping interface length in excess of 700 mm2 of glue. This appears to
be the most rigid joint with a reasonably high capacity.
Jacob Hammer Page 25 of 42 14/07/2017
Appendix
Jacob Hammer Page 26 of 42 14/07/2017
Test Strength (N) Displacement (mm) Glue Stress (MPa) Stick Stress (MPa)
Test Strength (N) Displacement (mm) Glue Stress (MPa) Stick Stress (MPa)
Test Strength (N) Displacement (mm) Glue Stress (MPa) Stick Stress (MPa)
Sample # Test Strength (N) Displacement (mm) L (mm) W (mm) Stress (MPa)
3 1580 1.580 28.05 30.5 1.847
1 2004 2.810 30.31 30.5 2.168
4 2194 2.194 28.07 30.5 2.563
2 3152 3.152 30.79 30.5 3.356
Frame Station P V2 V3 T M2 M3
Text mm N N N N-mm N-mm N-mm
Pier Left 0 -2500 0 0 0 0 0
Pier Left 75 -2500 0 0 0 0 0
Pier Left 150 -2500 0 0 0 0 0
Pier Right 0 -2500 7.276E-12 0 0 0 4.657E-10
Pier Right 75 -2500 7.276E-12 0 0 0 -8.004E-11
Pier Right 150 -2500 7.276E-12 0 0 0 -6.257E-10