Material Testing of M12 Bolts

This is part of “Impact Performance Research” of “Research into waste package

impact and fire accident performance”

Steve Burley and Qingming Li

School of Mechanical, Aerospace and Civil Engineering

The University of Manchester, Sackville Street
Manchester M13 9PL

i
This report has been prepared by The University of Manchester under contract to NDA. The
report has been reviewed by NDA, but the views expressed and conclusions drawn are those
of the authors and do not necessarily represent those of NDA.

Conditions of publication
This report is made available under the NDA Transparency Policy. In line with this policy,
the NDA is seeking to make information on its activities readily available, and to enable
interested parties to have access to and influence on its future programmes. The report may
be freely used for non-commercial purposes. However, all commercial uses, including
copying and re-publication, require permission from the NDA. All copyright, database rights
and other intellectual property rights reside with the NDA. Applications for permission to use
the report commercially should be made to the NDA Information Manager.
Although great care has been taken to ensure the accuracy and completeness of the
information contained in this publication, the NDA can not assume any responsibility for
consequences that may arise from its use by other parties.
© Nuclear Decommissioning Authority 2010. All rights reserved.

Bibliography
If you would like to see other reports available from NDA, a complete listing can be viewed
at our website www.nda.gov.uk, or please write to the Library at the address below.

Feedback
Readers are invited to provide feedback to the NDA on the contents, clarity and presentation
of this report and on the means of improving the range of NDA reports published. Feedback
should be addressed to:

John Dalton,
Head of Communications,
Nuclear Decommissioning Authority (Radioactive Waste Management Directorate),
Curie Avenue,
Harwell Campus,
Didcot,
Oxon,
OX11 0RH, UK

ii
Document Issue Record

Document Title: Material testing of M12 bolts

Project Reference: NDA-WT-150505-T2-Bolt-Test
Client: Radioactive Waste Management Division (RWMD), Nuclear
Decommissioning Authority (NDA)

Issue Description Drafted by Commented/modified by Approved by

1 Draft for internal SB QML QML
comment (24/10/2008) (13/11/2008) (13/11/2008)
2 Draft for client SB & QML Janak Patel (Rolls Royce)
comment (14/11/2008) Conrad Izatt (Arup)
Gordon Turner (NDA)
(03/09/2009)
3 Draft after considering SB QML QML
reviewers’ comments (31/01/2010) (18/02/2010) (19/02/2010)
4 Final report SB/QML Sarah Vines, Gordon QML
(15/05/2010) Turner (NDA) / SB, QML (17/05/2010)
(30/04/2010)

Note: SB (Steve Burley), QML (Qingming Li),

iii
Contents

Contents ...................................................................................................................................................iv
1. Introduction .........................................................................................................................................1
2. Testing...................................................................................................................................................2
2.1 Sample recovery .................................................................................................................. 3
2.2 Waisted Specimens ............................................................................................................. 3
2.2.1 Results .............................................................................................................................. 4
2.2.2 High strain rate equations ............................................................................................. 5
2.2.3 Comments on material properties............................................................................... 6
2.3 Bolts...................................................................................................................................... 7
2.3.1 Expected bolt properties ................................................................................................ 7
2.3.2 Results ............................................................................................................................... 9
2.3.3 Comparison with previous work.................................................................................10
2.4 Shear ...................................................................................................................................11
2.4.1 Results .............................................................................................................................12
2.4.2 Comments.......................................................................................................................12
3. Conclusions ........................................................................................................................................13
References...............................................................................................................................................14
Tables.......................................................................................................................................................15
Figures .....................................................................................................................................................20
Appendix 1 Calibration.........................................................................................................................40
A1. Testing machine load cell ...............................................................................................40
A2. Extensometer ...................................................................................................................40
A3. Micrometer .......................................................................................................................40
A4. Data Recorder ..................................................................................................................40
A5. LVDT ................................................................................................................................40
Appendix 2 Supplement .......................................................................................................................42
A2.1 True stress-strain curves of waisted specimens.........................................................42
A2.2 Bolt load-extension curves ...........................................................................................54
A2.3 Shear tests .......................................................................................................................64

iv
1. Introduction

The mission of the Nuclear Decommissioning Authority (NDA) is to deliver safe,

sustainable and publicly acceptable solutions to the challenge of nuclear clean-up and
waste management. This means never compromising on safety, or security, taking
full account of social and environmental responsibilities, always seeking value for
money for the tax payer, and actively engaging with stakeholders. The NDA Disposal
System Safety Case is being developed for the long-term management of higher
activity wastes. This includes safety assessments of transport to and operations at a
future repository.

As part of its work on safety, the NDA’s Radioactive Waste Management Division
(RWMD) is developing a generic Disposal System Safety Case (DSSC) to provide
evidence that transport to and operations at a Geological Disposal Facility (GDF)
meet all applicable regulatory requirements. This will look at all aspects of the GDF
and consider the consequences of any faults in a deep underground facility. This will
be assessed by the regulators before a licence to operate can be granted. A key issue
in the design and assessment of such waste packages is to provide safety against
impact accidents during transport to and handling at a future deep underground
repository.

To assess the consequences it is important to be able to estimate the amount of

material released at each fault location. One way of doing this is to determine what
fraction of the radioactive material is released from a waste package following a
defined accident. Release fractions have been determined using both physical tests
(e.g. drop tests on full scale waste containers containing simulated wasteforms, impact
tests on scaled physical models) and numerical analysis (e.g. computer-based finite
element models). The physical tests have provided a basic understanding of how
waste packages perform and the results of these tests have been used to develop,
refine and validate computer models. With a range of different container types, a
variety of wasteforms and numerous impact scenarios (e.g. different drop orientations
and target features), there are too many combinations to test the full range of waste
package responses. Computer models are therefore needed to allow the assessment of
all defined events for the different waste packages.

Numerical modelling, primarily but not exclusively finite element modelling (FE), is
used to model structures under both static and dynamic loading. Individual
components, such as lids, bolts, drums, contents, and the interactions between them,
can all be modelled. It is necessary to fully understand the response of all individual
components before progressing. This is usually at a material level, and thus, an
understanding of the mechanical properties of the component material is often
sufficient. It is important to have confidence in all parts of the modelling process and
a key feature in maintaining the integrity of the waste package is the performance of
the lid-body interface of the waste container. The current series of tests were
performed to improve confidence in the modelling of bolts in the waste containers.

In order to generate confidence in the models used, physical tests must be performed
and the results compared to the predictions of the model. The model should be
capable of predicting all critical failures. When there is confidence in the model,

1
designs can be changed and the effect of the change can be predicted by modelling
without the need to perform extensive experimental programmes.

An example is shown in Figure 1 of the external damage to a 500 litre drum. The lid
of the drum is attached by twelve 30 mm long M12 bolts. These bolts pass through
clearance holes in the lid into a threaded hole in the body flange. One possible release
pathway in an impact accident is the failure of one or more of these bolts leading to an
opening in the lid–body interface. As part of the model validation process, FE impact
modelling of a 500 litre drum has been performed [Arup (2007)] and compared to
drop tests onto a flat target of two full-scale 500 litre drum stimulant waste packages
in a lid-edge and an axis horizontal orientations from a height of 10 metres [Nirex
1989]. There was a discrepancy between the number of bolts that failed in the test
(seven) for the lid-edge orientation and the predicted failure in modelling (four).

Mechanical performances of bolts used in waste containers have been reported in

Arup (2002). The material data was obtained from a test programme on off-the-shelf
bolts to a similar specification. However, the specification allows considerable
variation in the physical properties. It has been suggested that the bolts whose
properties were tested and the bolts used in the drop tests may not have had similar
properties. The objective of this work is to determine the mechanical properties of the
bolts used in actual drop tests noting that the bolts in these drop tests could not be
located, but bolts from other 500 litre drum drop tests performed at the same time
were available and retrieved.

2. Testing

Tensile tests were performed on bolts and waisted specimens at nominal strain rates
of 0.001, 0.1 and 1/s.

An Instron 8500 hydraulic universal testing machine was used for all tests. The load
on the sample was measured using the built-in load cell. Crosshead displacement was
measured using the built-in transducer. Local sample deformation was measured in
one of two ways. For the waisted specimens, extension was measured using an 8 mm
gauge length extensometer (the small daylight between the chucks prevented the use
of an extensometer with a longer gauge length). For the bolts, extension was
measured using a linear variable differential transformer (LVDT). Calibration details
of the testing machine and instruments are shown in Appendix 1.

Data was recorded by two methods

1. Using the internal data capture equipment built into the firmware and software
that controls the testing machine. Capture is limited to about 8000 samples in
two channels. Overflowing the buffers causes the test to stop.

2. Using an external transient recorder to capture external outputs from the

testing machine controller. The output sensitivity (i.e. mechanical units per
volt) is controllable in software.

2
Because of the limited number of data available from the testing machine, the data
recorded externally was used for data processing.

Data was processed using the matrix manipulation and plotting package Matlab.

The 0.2% proof stress was determined by the method described in BS EN 10002 (for
tensile) and BS 6105 (for bolts). The method is similar in both cases. A straight line
satisfying a least squares fit to a portion of the stress-strain curve was determined.
The line was offset in the positive strain direction by 0.2% and the intersection point
of the new line and the curve was found. A sample calculation is shown graphically
in Figure 2.

Some of the data, primarily the LVDT data, was passed through a low-pass Gaussian
filter in Matlab. When applied to a sharp-fronted signal (e.g. a square wave), the
Gaussian filter maximises the rate-of-rise in the filtered signal without overshoot or
ringing. It has the disadvantage that its fall-off in response with increasing frequency
is relatively slow. A given filter applies a fixed time change to the data, independent
of frequency. This can be eliminated either by a passing a reversed copy of the
filtered signal through the filter or, if the data is being smoothed for graphical
purposes, passing both sets of data (e.g. stress and strain) through the filter.

2.1 Sample recovery

35 bolts were removed from previously tested 500 litre drums at NDA’s storage
facility on January 14 2008. The drums from which the bolts were taken were noted.
A check of the results showed no unexplained patterns indicating that all the bolts
were likely to have been provided from the same supplier and batch. The bolts were
M12 with a nominal shank length of 30 mm. The shank was threaded along its entire
length. Arup (2007) identifies the bolt material as A2-70.

2.2 Waisted Specimens

A number of bolts were machined into waisted specimens for testing. The shape is
shown in Figure 3. The diameters of the waisted specimens were measured using a
micrometer, and are shown in Table 1.

The specification for the bolts requires the threads to be produced by rolling [BNFL
(1989)]. The rolling process will deform the material around the threads up to an
unknown depth below the surface. It is possible that the properties of the worked
areas are different from the properties at the centre of the bolts.

Waisted specimens were tested at three nominal strain rates; 0.001, 0.1 and 1 per
second, shown in Table 2. The specimens were gripped in chucks with a loading
capacity of 20 kN, shown in Figure 4. The through holes eliminate any possibility of
bending at right angles to the axis of the hole. By arranging the holes in the pair of
chucks to be at right angle, all bending in the sample is eliminated, and the loading is
in pure tension.

3
2.2.1 Results

Results are summarised in Table 3. True stress-strain curves and corresponding strain
rate curves are shown in Appendix A2.1. A typical waisted specimen after failure is
shown in Figure 5. The graphs show the strain measured over the gauge length of the
extensometer. Necking and fracture did not usually occur within the gauge length on
the specimen. The elongation is calculated as the percentage change in length of the
specimen, which is measured after recovery. This causes a large difference between
elongation values in Table 3 and Appendix A2.1.

True stress and strain were calculated as

ε T = ln(1 + ε )
σ T = σ × (1 + ε )
where ε is the engineering strain (extension over the gauge length divided by the
original gauge length) and σ is the engineering stress (load divided by original area).
Plastic true strain was calculated by subtracting the elastic portion of the strain (i.e.
σ/E)

Figure 6 is a multi-plot showing the true stress versus the true strain (including the
elastic portion). The effect of increased strain rate on the stress-strain behaviour is
that the stress for any given strain increases as the strain rate increases. This
behaviour is expected.

The strain was measured using an 8mm gauge length extensometer. Almost all
failures happened outside the gauge length, and therefore, the maximum true plastic
strains in stress-strain curves in Appendix A2.1 are significantly smaller than the
percentage elongations in Table 3, which was obtained from the failed specimen.
Therefore, the percentage elongation given in Table 3 can be considered as an index
of material ductility, but the maximum true plastic strain in the stress-strain curves in
Appendix A2.1 cannot represent material ductility. Because various degrees of
necking happened in the gauge length, the extensometer measured various elongations
over the gauge length, which caused the different turning points seen in Fig.7(b).

Examination of the elastic region of the stress-strain curves shows that the relation is
often not linear. The measured modulus varies from about 150 to just over 200 GPa.
The reason for this variation is not clear. Typical reasons for this include:

1. Lack of linearity in the extensometer or measuring equipment. This was

checked and found to be acceptable.
2. Poor grip of the extensometer onto the specimen. This usually manifests as
step changes in the measured strain (as the extensometer grips move on the
sample), which was not seen here.
3. Noise from external sources (mains hum etc). This was not seen. The
duration of the tests precludes this.
4. Bending in the sample can arise when the samples are poorly seated in the
chuck. Care was taken to avoid this.

4
Several other causes for inconsistency come under the general heading of ‘operator
error’. The variation in tensile modulus between different batches of steel is expected
to be small. A few waisted specimens were machined from our stock material and
tested under conditions identical to those used here; these were found to show much
better consistency.

Many of the bolts were damaged in the drum impact tests. On recovery for this work,
several bolts were found to have damaged threads. An early decision was taken that
these bolts would be used to make the waisted specimens, leaving the bolts with clean
threads to be used on the whole bolt tests. Pre-existing plastic deformation in the
waisted specimens could be the cause of the inconsistency.

The variation in the modulus has a consequential effect on the variability in the
0.2% proof stress, but in terms of the energy absorbed this elastic energy is negligible
compared to the energy absorbed during plastic deformation.

2.2.2 High strain rate equations

The general effect of increasing strain rate on the tensile properties of steel is that the
yield point increases and the extension at failure decreases. However, the latter cannot
be consistently demonstrated based on the data in Table 3

Strain rate effects were modelled in Arup (2002) using the Cowper-Symonds relation,
i.e.
1
σY  ε
 P
= 1+  
Y C
where σY are Y are the dynamic and quasi-static representative stresses corresponding
to a given plastic strain, ε
is the nominal strain rate, and C and P are Cowper-
Symonds constants determined from tensile testing results at various strain rates.
Cowper-Symonds model only scale the representative stress, i.e. whole stress-strain
curve is shifted using the same value of the representative stress offset. Frequently,
yield stress (corresponding to 0.2% plastic strain in most cases) is taken as the
representative stress [Arup (2002)].

The model is only applicable to the plastic part of the total strain, so the plots for the
Cowper-Symonds models should strictly be shifted to the right to remove the elastic
strain.

In order to consider the material variation in the same testing group, true stress-strain
curves of four repeat tests at same nominal strain rate, as shown in Table 1, are
averaged to obtain an averaged true stress-strain curve at that strain rate, which are
presented in Figure 7. The curves are similar, which are consistent for all the bolts
being from the same batch. The strengths are above those required by the standard.

Figure 8 gives the averaged true stress-strain curves from the present tensile tests on
waisted specimens at three nominal strain rates, i.e. 10-3 1/s (quasi-static), 0.1 1/s and
1.0 1/s. Yield stress is taken as the 0.2% true proof stress, i.e. Y= 666MPa. The best
data fitting of Cowper-Symonds constants to the dynamic 0.2% true proof stresses

5
leads to C=58.0 1/s and P=1.86. Since Arup (2002) noted (p 23) that the strain rates
in the modelling were up to 50/s, the predicted true stress-strain curves based on the
present averaged quasi-static true stress-strain curve and Cowper-Symonds relation at
strain rates of 0.1, 1.0 and 50 1/s are presented in Figure 8.

Figure 9 shows the predicted true stress-strain curves based on the present averaged
quasi-static true stress-strain curve and Cowper-Symonds constants (C=300 and
P=1.7) in Arup (2002) at 0.1, 1.0 and 50 1/s strain rates.

It should be noted that the Arup (2002) test programme was optimised on A4-70 bolts
and only conducted tensile tests on full threaded A2-70 M16 bolts. The quasi-static
stress-strain data of full threaded A2-70 M16 bolts from Arup (2002) (p 23) is shown
in Table 4. Tensile test results of threaded bolts and waisted specimens of A4-50
M16 bolts were compared in Fig.3.18 in Arup (2002) at quasi-static strain rate (10-3
1/s), which is reproduced in Figure 10. It shows that the ductility of the waisted
specimen is much greater than the ductility of the threaded bolt while a reduction of
the yield stress is observed for the waisted specimen. If such difference exists for A2-
70 M12 bolts, the tensile tests on full threaded A2-70 M16 bolts in Arup (2002) may
not be treated as material tests for A2-70 stainless steel. Therefore, comparisons and
predictions in Figure 8 and Figure 9 are all based on the present tensile testing results
on waisted specimens.

According to Figure 8 and Figure 9, present prediction (i.e. the predicted true stress-
strain curve based on Cowper-Symonds constants from the present study) is
considerably larger than Arup’s corresponding prediction [i.e. the prediction based on
Cowper-Symonds constants from Arup (2002)] in the realistic range of strain-rates. At
50 1/s strain rate, the Cowper-Symonds factor based on 0.2% proof true stress in the
1 1
 ε
 P  50 1.86
present study is 1 +   = 1+   = 1.92 , which can be compared to the
C  58 
1
 50 1.70
corresponding value of 1 +   = 1.35 based on Cowper-Symonds constants
 300 
from Arup (2002).

Figure 8 and Figure 9 show discrepancy between the measured and the predicted true
stress-strain curves in the majority of plastic strain experienced by the tested specimen.
This implies that Cowper-Symonds constants obtained by fitting the dynamic 0.2%
proof stresses at tested strain rates may not be representative. It was found that good
fits were obtained to the entire stress-strain curves at strain rates of 0.1 and 1.0 1/s by
using the true stress at 1%-3% true plastic strain as the representative stress in
Cowper-Symonds relation, as shown in Figure 11. The fitted Cowper-Symonds
constants are given in Table 5. Their relative differences at low strain rates are very
small. Even at 50 1/s strain rate, their relative differences are less than 5%.

2.2.3 Comments on material properties

The elastic-plastic properties of the material can be given with quite sufficient
accuracy using a bilinear model. Curvature in the stress-strain curve is small. There

6
is some curvature around the yield point, but this is insignificant in the overall
response.

The percentage elongation of the tested samples was in the range of 15–25%, as
shown in Table 3. This measure is an artefact of the tensile test due to the occurrence
of necking, and is not a material property. However, given enough knowledge of the
material and the testing machine, the behaviour of the specimen can be modelled
numerically. Agreement between the modeling and the test data can be taken as a
validation of the material model.

The strain rate effect on stress-strain curve can be estimated based on Cowper-
Symonds relation. However, the determination of constants C and P depends largely
on the testing range of strain rates and the selection of the representative flow stress.
Extrapolation to higher strain rates should be taken with care, especially when large
amounts of material experience higher strain rates.

The present and previous test data relate to specimens with different geometric
dimensions. In general, the length-to-diameter ratios in the current series of tests
(both tension and shear) are smaller than those used in previous work. Three-
dimensional effects can significantly influence the overall response of the sample.
Without further numerical investigation, which is not in the scope of this project, such
effects cannot be quantified. Although the influences of manufacturing and damage
factors on the different bolt behaviours between the current and previous tests cannot
be excluded, geometrical effects may play more important roles than other factors.

2.3 Bolts

Bolts were tested in the jig shown in Figure 12 at three nominal strain rates described
in Table 6. The bolt under test was inserted into the upper part and prevented from
rotating by a grub screw. The lower part was threaded onto the bolt and the parts
hand tightened. The assembly was gripped by the jaws in the testing machine and
pulled apart. The relative displacement of the two parts was measured using a linear
variable differential transformer (LVDT). Load and cross-head displacement were
measured by internally-fitted transducers, respectively. All data were recorded on a
Nicolet ‘Multipro’ transient recorder.

LVDTs are inherently noisy devices and their output must be filtered at some stage,
either in the control unit or in post-processing. A single-pass Gaussian filter was
applied to the data in post-processing.

2.3.1 Expected bolt properties

The bolts are 12 mm metric coarse thread made from BS6105 A2-70 stainless steel.
[BNFL(1989)]. The material properties are listed in the standard and reproduced as
Table 7. The thread properties are given in Table 8. Using the values above, the peak
load on a bolt should exceed

84.3 × 700 N = 59 kN

and the 0.2% proof stress should exceed

7
 84.3 × 450 N = 38 kN

The bolts are not long enough for the extension property to apply, but for
completeness, the extension should be greater than

0.4 × 12 = 4.8 mm
which corresponds to a percentage elongation of

4 .8
= 30%
15.8

over a 15.8 mm gauge length.

The stiffness of bolts in joints is considerably less than what would be expected from
a simple analysis based on the cross-section and length of the shank. Deformation in
the head and at the threads both act to reduce stiffness. Accurate prediction of the
compliance is complicated by the effects of the size of the head, the shape of the head,
the flatness of bearing surfaces, and the tolerances in thread forming. Short bolts with
a length less than about four diameters are known to be unusually problematic
because the effects of the head and nut deformations dominate the overall stiffness of
the bolt. ESDU 85021 provides a method to estimate the expected stiffness of a bolt.
The method breaks the system into three parts, i.e. the head and shank, the threaded
part and the nut.

The effect of the bolt head is expressed as an extension to the effective length of the
shank. There are two parts in the extension. The first allows for the diffusion of the
load into the shank and can be taken as 10-15% of the shank diameter. The second
allows for deformation in the head, which, for a standard hex-headed bolt,
corresponds to 15-20% of the shank diameter. However, the heads in the bolts used
here were non-standard, with a domed head and a recess, both of which act to reduce
stiffness, and therefore, the effective extension can be expected to be more.

The effective diameter of the threaded part is taken as the root diameter plus 0.3 times
the thread depth.

The effect of elasticity in the thread/nut interface and the nut is expressed as an
extension to the threaded part. The extension is about 1.5 to 3 times the effective
diameter of the threaded part for a standard nut. The threaded jig used here is stiffer
than a standard nut.

The compliance is given by the sum of the compliance of these three parts, i.e.

S = S h + St + S n

where Sh is the compliance of the head, St is the compliance of the threaded part, and
Sn is the compliance of the thread/nut interface and nut, or

8
 1 1.15 × 1.2 × d  + 3d e 
S=  + 
E  πd 2 / 4 πd e2 / 4 

where E is Young’s modulus (190 GPa), d is the nominal bolt diameter (12 mm), ℓ is
the nominal length (15.8 mm), and de is the effective diameter of the threaded part
[10.106 + 0.3×(12-10.106) mm = 10.67 mm]. Numerically,

1.15 × 1.2 × d
= 0.146
πd 2 / 4

= 0.176
πd e2 / 4
3d e
= 0.358
πd e2 / 4

S=
(0.146 + 0.176 + 0.358) × 10 3 = 3.58 × 10 −9 m/N
9
190 × 10

giving the compliance for the bolt as about 3.6 µm/kN. In addition, the bolt head
bears on a small area of the upper part of jig. Treating this bearing area as the area
difference between the area of the hex head (19 mm across the flats) and the shank
area and taking length equal to ℓ give compliance of this part to be about 0.4 µm/kN.

Hence the expected compliance of the rig/bolt system is about 4.0 µm/kN. This
estimate may be high because the part of the rig with an internal thread will be stiffer
than a standard nut, but may be low because of any extra stiffness arising from pre-
tension in the bolt. The compliance of the threaded part is about 1 µm/kN, thus about
75% of the observed deformation takes place outside the threaded part.

This represents a limitation to the data gathering process, in that the extension of the
bolt is not measured. It is thus not possible to present meaningful stress-strain curves.
The length of the active part of the bolt (15.8 mm) and the need to fully support the
head of the bolt make it impractical to use extensometer techniques to measure the
extension of the bolt. In principle, it would be possible to attach one or more strain
gauges to the bolt to measure the surface strain, but care would have to be taken to
eliminate the effects of the threads. The surface strain is not necessarily a good
measure of the average longitudinal strain.

2.3.2 Results

A typical failed bolt specimen is shown in Figure 13. The measured load-extension
curves are plotted in Figure 14 and in Appendix A2.2. Each plot includes the
measured maximum load, the computed 0.2% proof load, and the measured
compliance. Most bolts failed at an extension of 4–5 mm. This represents a strain on
the loaded section (15.8 mm long) of 25–30%.

Table 9 shows that the measured compliance of the bolt/rig was close to that expected.
The measured 0.2% bolt proof load, at 61–65 kN, was above the minimum required

9
by the standard (38 kN). There was no significant variation of proof load with
varying loading rate.

Table 9 shows that the measured bolt peak load, at 70–75 kN, was above that required
by the standard (59 kN). There is a very small increase in peak load with increasing
load rate except test b15, which has the highest peak load of all bolts, despite being
loaded at a rate of 0.001/s.

Table 9 shows a reduction in the strain at failure as the loading rate increased. The
tests at a strain rate of 1/s failed at a strain of 26-7%, whereas the tests at a rate of
0.001/s failed at 32-3%. The failure strains of the slow rate tests exceed the
requirements of the standard (30%) but those at 1/s do not. However, the length of
the bolts is shorter than that required for the test to be valid.

The compliances vary from about 3.3 to 4.4 µm/kN. This is in agreement with the
expected values.

2.3.3 Comparison with previous work

The previous bolt testing work summarised in Arup (2002) (Chapter 3) contains
several stress-strain graphs for bolts. The curves in that report are plots of

F
(1 + ε ) versus ln (1 + ε )
Ar
where ε is the strain defined as extension divided by nominal length, F is the applied
force, and Ar is the root area. The extension of the bolts was taken as the
displacement of the crosshead of the testing machine and the nominal length was
25 mm [Arup (2002) p 12]. The root area is assumed to be the area of the minor
diameter of the bolt thread. At small strains the extension is dominated by
compliance in the testing rig, hence the measure of strain is only approximate. The
work tested bolts of diameter M14, M16 and M18. It was concluded that size had
only a small effect on the stress-strain curves of the bolts, see Arup (2002) (p 14) and
Figures 3.19 and 3.20.

Figure 15 is a copy of Figure 3.1 from the report. The figure shows the result for
three A2-70 M16 bolts tested at nominal strain rates of 0.001, 0.1, and 4.0/s [tests 165,
167, 168 respectively, see Arup (2002) pp 22–3].

For comparison, Figure 16 is a plot of stress versus strain, as defined above and using
the extension as measured with the LVDT, for all bolt tests. Figure 17 is a similar
plot using the crosshead displacement to measure displacement. Both figures have
been adjusted (by introducing an offset into the displacement data) so that all curves
cross around 300 MPa. The only major difference between the two figures is in the
initial elastic portion of the curves, where the strain as measured by the LVDT is less
than that calculated using the crosshead displacement. This is due to compliance in
the jig and testing machine. The ratio is a little under 2:1.

Figure 18 is a combination of Figure 15 and Figure 17. These two figures were
imported into an image processing package and resized such that the axes of the two
figures were at the same scale. The measured yield point is a little lower for the

10
previous work, but the primary difference is in the post-yield behaviour. The bolts in
the previous work showed considerable work hardening, whereas the bolts tested here
showed softening features. The failure strains are similar.

As noted above, joints are not simple structures, even for bolts in tension. At the head
end, forces at the interface between the bolt and the mated part interface are
transferred via the head to the shank. The stress in the shank will not be uniform
across the diameter close to the head. Similarly, load is transferred to the bolt via the
external threads in contact with an internally threaded part (usually a nut, but the
flange in the Drums, and part of the rig in these tests). Close to the nut (etc) stresses
are higher in the outer part of the cross-section of the bolt. The threads in this part
will act as stress raisers. Both phenomena mean that the behaviour of short bolts is
dominated by load transfer effects and can differ greatly to that of an otherwise
identical longer bolt. This may explain the differences between the two sets of data.

The areas under all the curves in Figure 18 are approximately similar, indicating a
similar work to failure.

2.4 Shear

Bolts were tested in the jig shown in Figure 19 at an extension rate of 1.2 mm/s. The
jig, which was made of high strength steel (FORA 400, yield strength of 1100 MPa,
UTS of 1300 MPa), together with the bolt assembly, was gripped in the jaws of the
testing machine (Instron 8500) and pulled apart. The relative displacement (extension)
of the two parts between A and B in Figure 19 was measured using a LVDT. Load
and cross-head displacement were measured by internal fitted transducers. All data
were recorded on a ‘Nicolet’ transient recorder. LVDTs are inherently noisy devices,
and therefore, their output must be filtered either in the control unit or in post-
processing. In the present data analysis, a single-pass Gaussian filter was applied to
the data in post-processing.

If a headed bolt is tested, the upper part of the bolt head bears on the jig and applies
tensile forces onto the shank. These forces make the test asymmetric and can damage
the rig, especially after yielding occurs. In order to avoid these problems, the heads of
the bolts were removed. The shank to be tested was passed through the holes in the
two parts of the rig. Friction held the shank in place at the beginning of the test; as
the test progressed the blade deformed the central part of the shank, holding the
specimen firmly in place.

Nominal shear stress is defined as

F
τn =
2A
where F is the tensile load obtained from the load transducer in the loading machine
and A is the cross sectional area at the root of bolt. It should be pointed out that the
F
shear stress definition used in Arup(2002) is τ = , which does not truly represent the
A
shear stress on the cross-sectional surface of a bolt because the pulling load is shared
by two shear surfaces. The standard shear stress definition will be used in this report
except in Table 11 and Figure 23 and Figure 24 where Arup(2002)’s definition of

11
shear stress is used. The non-standard definition does not affect the usefulness of
comparisons of numerical models to experimental data.

Nominal shear strain is defined as

∆
γn =
D
where ∆ is the extension measured by the LVDT and D=12 mm is the diameter of the
bolt. Nominal shear strain rate is defined by
∆

γ
n =
D
where ∆
= 1.2 mm/s is the extension rate, which gives a nominal shear strain-rate of
0.1/s.

2.4.1 Results
Typical failed specimens are shown in Figure 20. The measured load-extension
curves are plotted in Figure 48 to Figure 52. Each plot includes the measured
maximum load. It shows that most bolts reached a peak load of about 80 kN at an
extension of 2–3 mm. The results are summarised in Table 10. Results were
reasonably consistent, with a standard deviation of about 3% of the average value.

2.4.2 Comments

Previous bolt testing work summarised in Chapter 3 of Arup (2002) contains several
examples of bolt shear tests. The tests were fundamentally different in that the
thickness of the inner shearing blade was 30 mm for the tested bolts with larger
diameters of 16 and 18 mm, which was much larger than the 10 mm inner blade
thickness used here for 12 mm diameter bolts. Examination of the text and the
magnitude of the maximum reported shear stresses (see p 2 and Figs.3.6–8 and 3.17 in
Arup (2002)) suggests that the nominal shear stress in Arup (2002) was calculated as
applied load divided by the root area of the thread (Note that the load is applied to two
shear planes).

Figure 3.17 in Arup (2002) is reproduced as Figure 23 in the present report. Both sets
of data are plotted in Figure 24. The maximum stresses reached in the present work
are considerably lower than those reported in Arup (2002). Unfortunately, only two
tests were performed on A2-70 steel bolts in Arup (2002), i.e. tests Y0W166A & B in
Table 3.7a, test 166 in Table 3.8 and Figure 3.17. Table 3.8 indicates a 0.2% shear
stress of over 740 MPa and an ultimate engineering shear stress of over 1100 MPa.
The summary data is reproduced as Table 11 below.

For double shear tests such as those performed here, the load is introduced into the
bolts by contact between the bolt and the blades. At any point in the bolt, the stress is
due to a combination of the effects of the interactions at either side of the inner blade.
The stress distribution will vary with the distance from the faces of the inner blade,
and the stress distribution will be a function of blade thickness. Hence, tests on short
bolts give different results to tests on long bolts of the same diameter. It is possible
that the differences seen in Figure 24 are a consequence of this.

12
3. Conclusions

A2-70 M12 bolts recovered from previous 500 litre drum drop tests have been
successfully recovered. The bolts were tested in tensile and shear loadings under both
static and dynamic conditions. Additional bolts were machined down to provide
standard test specimens for material tensile testing.

The 0.2% proof stress and maximum load for the bolts were within the specification
of the tested bolts. The elastic-plastic properties of the material were consistent such
that the performance can be given with quite sufficient accuracy using a bilinear
model.

The Cowper-Symonds model used in Arup (2002) is capable of fitting to the

measured data over the experimental measurements of strain and strain rate if proper
representative stress is selected. Extrapolating to higher strain rates (i.e. significantly
above 1/s) leads to a significant increase in the predicted yield stress, which needs to
be justified. If critical parts of the model are loaded at high strain rates, then the use
of the Cowper-Symonds constants determined at low strain rates could be a cause of
uncertainty.

The elastic compliance of the bolt was found to be in the range of 3.3–4.4 µm/kN in
the as-tested configuration. This was intended to be close to the in-service
configuration.

For the double shear loading tests the measured load-extension curves showed that
most bolts reached a peak load of about 80 kN at an extension of 2–3 mm.

The load is introduced into the bolts by contact between the bolt and the blades. At
any point in the bolt, the stress is due to a combination of the effects of the
interactions at either side of the inner blade. The stress distribution will vary with the
distance from the faces of the inner blade, and the stress distribution will be a function
of blade thickness. Hence, tests on short bolts give different results to tests on long
bolts of the same diameter. It is possible that the higher peak load and shorter
elongation to failure reported in Arup (2002) are a consequence of this.

13
References
Arup (2002). 3m³ Box Lid Bolt Investigation (report number 57718/07/03).

Arup (2007), Holistic methodology validation - Stage 1: validation of modelling of

wasteform global behaviour (Draft-1).

BNFL(1989), 3/PR1081372 “EP1 Product container lid bolt details” drawing, Risley,
Warrington, Cheshire.

BS EN 10002-1:2001 Metallic materials — Tensile testing — Part 1: Method of test

at ambient temperature.

BS 6105:1981. Specification for Corrosion-resistant stainless steel fasteners.

Matlab, The Mathworks Inc. Version 7.5.0.342 (R2007b), 2007.

Nirex (1989), Impact Behaviour of 500 Litre Intermediate Level Waste Drums, Nirex
Report No.70.

14
Tables

test Head Centre Free end

dia (mm) dia (mm) dia (mm) dia (mm) dia (mm) dia (mm)
T02 4.97 4.97 4.96 4.97 4.97 4.98
T03 4.99 5.00 5.01 4.99 5.00 5.01
T15 4.98 4.98 4.99 4.98 4.98 4.99
T16 5.02 5.01 5.01 5.01 5.01 5.01
T04 5.00 5.00 5.01 5.00 5.00 5.00
T05 4.99 4.99 4.99 4.99 4.99 4.99
T17 5.00 5.00 5.00 5.01 5.01 5.00
T18 5.03 5.03 5.03 5.03 5.03 5.03
T06 5.01 4.97 4.95 5.01 4.97 4.96
T07 5.00 4.99 4.98 5.00 4.99 4.98
T19 4.96 4.96 4.96 4.96 4.97 4.96
T20 4.98 4.98 4.97 4.98 4.98 4.98
Table 1 Waisted specimen diameters

Nominal strain Cross-head speed Data recording

rate (/s) (mm/min) interval (µs)
0.001 1.5 105
0.1 150 102
1 1500 10
Table 2 Tensile test parameters

15
Test Loading Modulus 0.2% Maximum Maximum Percentage
rate (/s) (GPa) engineering engineering true stress Elongation
proof stress stress (MPa) (%)
(MPa) (MPa)
T02 0.001 153 643 731 773 24
T03 0.001 189 641 718 754 19
T15 0.001 155 662 733 766 20
T16 0.001 206 703 740 818 25
T04 0.1 153 716 770 792 15
T05 0.1 151 609 748 782 19
T17 0.1 204 703 753 793 15
T18 0.1 175 708 755 788 15
T06 1 181 729 792 825 17
T07 1 168 753 805 856 20
T19 1 194 736 782 818 15
T20 1 199 721 796 849 17
Table 3 Summary results of tensile tests

A2-70
Yield stress 600 MPa
True 600 720 780 820 890 930 950 970
stress
(MPa)
Plastic 0 0.005 0.018 0.041 0.097 0.147 0.197 0.247
strain
C (s-1) 300
P 1.7
Table 4 Stress-strain data for A2-70 [Arup (2002), p 23]

Plastic strain 1% 2% 3%
True stress (MPa) 730 743 750
at 0.001 1/s
True stress (MPa) 743 769 776
at 0.1 1/s
True stress (MPa) 750 800 809
at 1.0 1/s
C (1/s) 1903 1733 1032
P 2.98 2.90 2.73
Table 5 Cowper-Symonds constants obtained from different representative stresses

16
 Nominal strain Cross-head speed Data recording Filter frequency
rate (/s) (mm/min) interval (µs) (3dB, Hz)
0.001 0.6 105 2
0.1 60 104 200
1 600 100 2000
Table 6 Bolt test parameters

Group Grade Property Diameter Tensile Stress at Extension

class range strength 0.2% AL min.‡
Rm, min. permanent
(MPa)* strain
Rp0,2
min.
(MPa)†
Austenitic A2 70 ≤M20 700 450 0.4d
Table 7 Properties of A2-70 stainless steel (BS6105:1981)
*Calculated as the peak load divided by the nominal tensile stress area of the thread.
†This test is applicable only to fasteners of lengths equal to twice the diameter (2d) or
longer.
‡This test is applicable only to fasteners with lengths equal to three times the diameter
(3d) or longer.

Nominal Pitch (mm) Basic minor Stress area Root area

(mm) diameter, d3 (mm2) (mm2)†
(mm)
BS 3692:1967, BS 3643- BS 3643-
p3 1:1981, p 1:1981, p 8
BS 3692:1967,
p 2.
BS 6105:1981,
p2
12 1.75 10.106 84.3 76.2
Table 8 Properties of 12 mm bolts
†taken as the area of the minor diameter (d1=9.85 mm) see BS 3643-1, pp 1, 2, 4, 48.

17
 Test Loading Compliance 0.2% proof Maximum Failure
rate (/s) (µm/kN) load (kN) load (kN) strain (%)
b06 0.001 4.3 61 70 32
b07 0.001 3.9 62 70 33
b15 0.001 3.7 64 75 32
b08 0.1 4.4 63 71 28
b09 0.1 3.8 63 73 28
b10 0.1 3.5 64 71 27
b11 1 3.7 63 73 27
b12 1 3.5 63 74 26
b13 1 3.3 65 73 26
b14 1 3.5 61 73 26
Table 9 Summary results of bolt tests (Note: Failure is defined as when the load drops
to 20 kN; failure strain is defined as the percentage elongation)

18
 Testing No Nominal Maximum Nominal Maximum
shear strain- load (kN) shear stress at nominal shear
rate (1/s) 0.2% nominal stress (MPa)
shear strain
(MPa)
ST2 0.1 84.3 322 500
ST3 0.1 83.9 308 498
ST4 0.1 82.4 307 489
ST5 0.1 81.0 323 480
ST6 0.1 81.2 318 482
SD 2 10 10
Table 10 Shear test results

Test ref Nominal Maximum Nominal Maximum

shear strain- load (kN) shear stress at nominal shear
rate (1/s) 0.2% nominal stress (MPa)
shear strain
(MPa)
166 0.001 (not given) 745 1147
Table 11 Shear tests results from Table 3.8 in Arup (2002) for 16 mm A2-70 bolts

19
Figures

Figure 1 An example of 500 litre waste package drum after testing

20
 8
x 10
8

6 proof stress 659 MPa

4
stress

1 stress−strain data
data used for slope
fitted slope
0 offset slope
zero crossing
proof stress
−1
0 0.001 0.002 0.003 0.004 0.005 0.006 0.007 0.008 0.009 0.01
strain
Figure 2 An example of the calculation of 0.2% proof stress

Figure 3 Waisted specimen for tensile test

21
Figure 4 Waisted specimen and chucks

Figure 5 A waisted specimen after failure

22
 true stress−strain
Filtered at (rate * 50) Hz and aligned at 100 MPa.
900

800

700

600

500
true stress (MPa)

400

300

200

100

−100
−0.01 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09
true strain
Figure 6 Multi-plot of the total (i.e elastic and plastic) true stress-strain curves for
tensile tests. (Red curves – strain rate 0.001 1/s; green curves – strain rate 0.1 1/s;
blue curves – strain rate 1.0 1/s.)

23
 Average stresses: rate=0.001
1200

ave stress (0.9−1.1%) 730 MPa

ave 0.2% proof stress 666 MPa

1000

800
true stress (MPa)

600

400

200
T02
T03
T15
0 T16
average

−0.01 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09
plastic true strain

Figure 7 (a) Individual and averaged true stress-strain curves from present tests at
0.001 1/s (quasi-static) strain rate.

24
 Average stresses: rate=0.1
1200

ave stress (0.9−1.1%) 756 MPa

ave 0.2% proof stress 688 MPa

1000

800
true stress (MPa)

600

400

200
T04
T05
T17
0 T18
average

−0.01 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09
plastic true strain

Figure 7 (b) Individual and averaged true stress-strain curves from present tests at 0.1
1/s strain rate.

25
 Average stresses: rate=1
1200

ave stress (0.9−1.1%) 788 MPa

ave 0.2% proof stress 742 MPa

1000

800
true stress (MPa)

600

400

200
T06
T07
T19
0 T20
average

−0.01 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09
plastic true strain

Figure 7 (c) Individual and averaged true stress-strain curves from present tests at 1.0
1/s strain rate.

26
 Comparison of measured and predicted high strain rate properties
using 0.2% proof stress C= 58 P=1.86
1500

1000
true stress (MPa)

500

average 0.001
58/1.86/0.1
average 0.1
58/1.86/1
0 average 1
58/1.86/50

−0.01 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09
plastic true strain

Figure 8 Tested and predicted true stress-strain curves based on Cowper-Symonds

constants from present tests and using 0.2% proof stress as the representative stress

27
 Comparison of measured and predicted high strain rate properties
using previous values C= 300 P=1.70
1200

1000

800
true stress (MPa)

600

400

200
average 0.001
300/1.70/0.1
average 0.1
300/1.70/1
0 average 1
300/1.70/50

−0.01 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09
plastic true strain

Figure 9 Tested and predicted true stress-strain curves based on Cowper-Symonds

constants from Arup (2002) and using 0.2% proof stress as the representative stress

28
Figure 10 Reproduction of Figure 3.18 in Arup (2002) to show the shank/thread
effects in tensile tests for A4-50 at 0.001 1/s strain rate

29
 Comparison of measured and predicted high strain rate properties
using ave stress at strains (0.9−1.1%) C=1903 P=2.98
1200

1000

800
true stress (MPa)

600

400

200
average 0.001
1903/2.98/0.1
average 0.1
1903/2.98/1
0 average 1
1903/2.98/50

−0.01 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09
plastic true strain

Figure 11 (a) Tested and predicted true stress-strain curves based on Cowper-
Symonds constants from present tests and using stress at 1% plastic strain as the
representative stress

30
 Comparison of measured and predicted high strain rate properties
using ave stress at strains (1.8−2.2%) C=1733 P=2.90
1200

1000

800
true stress (MPa)

600

400

200
average 0.001
1733/2.90/0.1
average 0.1
1733/2.90/1
0 average 1
1733/2.90/50

−0.01 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09
plastic true strain

Figure 11 (b) Tested and predicted true stress-strain curves based on Cowper-
Symonds constants from present tests and using stress at 2% plastic strain as the
representative stress.

31
 Comparison of measured and predicted high strain rate properties
using ave stress at strains (2.7−3.3%) C=1032 P=2.73
1200

1000

800
true stress (MPa)

600

400

200
average 0.001
1032/2.73/0.1
average 0.1
1032/2.73/1
0 average 1
1032/2.73/50

−0.01 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09
plastic true strain

Figure 11 (c) Tested and predicted true stress-strain curves based on Cowper-
Symonds constants from present tests and using stress at 3% plastic strain as the
representative stress.

Figure 12 Bolt test jig

32
Figure 13 Bolt after failure

load−extension for bolts

Filtered at (rate * 200) Hz and aligned at 10kN.

80
b06 (0.001/s)
b07 (0.001/s)
b15 (0.001/s)
70
b08 (0.1/s)
b09 (0.1/s)
b10 (0.1/s)
60 b11 (1/s)
b12 (1/s)
b13 (1/s)
b14 (1/s)
50
load (kN)

40

30

20

10

0
0 1 2 3 4 5 6 7 8
extension (mm)
Figure 14 Load-extension for bolts

33
Figure 15 Copy of Figure 3.1 from Arup (2002)

stress−strain (see text for definitions) for bolts

Filtered at (rate * 200) Hz and aligned at 300 MPa.
Strain measure: LVDT.
1500
b06 (0.001/s)
b07 (0.001/s)
b15 (0.001/s)
b08 (0.1/s)
b09 (0.1/s)
1200 b10 (0.1/s)
b11 (1/s)
b12 (1/s)
b13 (1/s)
b14 (1/s)
900
stress (MPa)

600

300

0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
strain
Figure 16 Stress-strain curve for bolts (LVDT)

34
 stress−strain (see text for definitions) for bolts
Filtered at (rate * 200) Hz and aligned at 300 MPa.
Strain measure: crosshead.
1500
b06 (0.001/s)
b07 (0.001/s)
b15 (0.001/s)
b08 (0.1/s)
b09 (0.1/s)
1200 b10 (0.1/s)
b11 (1/s)
b12 (1/s)
b13 (1/s)
b14 (1/s)
900
stress (MPa)

600

300

0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
strain
Figure 17 Stress-strain curve for bolts (crosshead)

Figure 18 Comparison of this work (coloured lines) and previous work (grey)

35
A

Figure 19 Shear jig

Figure 20 Selected bolts after shear testing (ST4 upper, ST5 lower)

36
 load−extension for bolts in shear
Filtered at (rate * 200) Hz and aligned at 10kN.

90
st2 (0.1/s)
st3 (0.1/s)
80 st4 (0.1/s)
st5 (0.1/s)
st6 (0.1/s)
70

60

50
load (kN)

40

30

20

10

0
0 1 2 3 4 5 6 7 8
extension (mm)
Figure 21 Load-extension curves for all shear tests

37
 Stress−strain for bolts in shear (area=CSA*2)
Filtered at (rate * 200) Hz and aligned at 10kN.

600
st2 (0.1/s)
st3 (0.1/s)
st4 (0.1/s)
st5 (0.1/s)
500 st6 (0.1/s)

400
nominal stress (MPa)

300

200

100

0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
nominal strain

Figure 22 Nominal shear stress-strain curves for all shear tests.

Figure 23 Shear testing results from Arup (2002) [Note: shear stress defined as load
divided by root cross-sectional area; shear strain defined as crosshead displacement
divided by root diameter]

38
Figure 24 Comparison of the present work (coloured curves) and the previous one
(grey) in Fig.5 [Arup (2002)]. Note: we used the same shear stress and strain
definitions as those used in Arup(2002), i.e., shear stress: load divided by root cross-
sectional area; shear strain: crosshead displacement divided by root diameter.

39
Appendix 1 Calibration

A1. Testing machine load cell

Calibrated yearly by Instron.

A2. Extensometer

Epsilon type 3442-008M-050-ST, which was calibrated from manufacture. The

extensometer was new at the start of this work.

A3. Micrometer

Instrument 3962469, calibrated by Cromwell tools 4 June 2008, cert 164226.

A4. Data Recorder

Nicolet (self calibrating).

A5. LVDT

The linear variable differential transformer (LVDT) used in the bolt extension tests
was calibrated using a micrometer head mounted on a robust frame. The output
voltage was measured for a series of displacements and a straight line least squares fit
performed on the data.

The data is given in Table 12 and plotted in Figure 25. The standard deviation of the
fit is about 12 µm, equivalent to the resolution of the voltage measurement.

displacement voltage
(mm) (V) fit data
5 0.500 10.05106
4 0.401 -0.02741
3 0.300
2 0.201
1 0.102
0 0.004
-1 -0.095
-2 -0.195
-3 -0.296
-4 -0.397
-5 -0.495
Table 12 LVDT calibration

40
 6
raw data
fit
standard deviation * 100

2 average standard deviation is 12 microns

displacement (mm)

−2

−4

−6
−0.5 −0.4 −0.3 −0.2 −0.1 0 0.1 0.2 0.3 0.4 0.5
outout voltage (V)
Figure 25 LVDT calibration

41
Appendix 2 Supplement

A2.1 True stress-strain curves of waisted specimens

8 T02
x 10
9

6
true stress (Pa)

5
Max true stress 773 MPa
4

−1
−0.01 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09
true strain

−4
x 10
15

10
true strainrate (/s)

−5
−0.01 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09
true strain

Figure 26 True stress and true strain rate for test T02

42
 8 T03
x 10
9

6
true stress (Pa)

5
Max true stress 754 MPa
4

−1
−0.01 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09
true strain

−4
x 10
15

10
true strainrate (/s)

−5
−0.01 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09
true strain

Figure 27 True stress and true strain rate for test T03

43
 8 T15
x 10
9

6
true stress (Pa)

5
Max true stress 766 MPa
4

−1
−0.01 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09
true strain

−4
x 10
15

10
true strainrate (/s)

−5
−0.01 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09
true strain

Figure 28 True stress and true strain rate for test T15

44
 8 T16
x 10
9

6
true stress (Pa)

5 Max true stress 818 MPa

−1
−0.01 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09
true strain

−4
x 10
15

10
true strainrate (/s)

−5
−0.01 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09
true strain

Figure 29 True stress and true strain rate for test T16

45
 8 T04
x 10
9

6
true stress (Pa)

5 Max true stress 792 MPa

−1
−0.01 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09
true strain

0.15

0.1
true strainrate (/s)

0.05

−0.05
−0.01 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09
true strain

Figure 30 True stress and true strain rate for test T04

46
 8 T05
x 10
9

6
true stress (Pa)

4 Max true stress 782 MPa

−1
−0.01 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09
true strain

0.15

0.1
true strainrate (/s)

0.05

−0.05
−0.01 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09
true strain

Figure 31 True stress and true strain rate for test T05

47
 8 T17
x 10
9

6
true stress (Pa)

5 Max true stress 793 MPa

−1
−0.01 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09
true strain

0.15

0.1
true strainrate (/s)

0.05

−0.05
−0.01 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09
true strain

Figure 32 True stress and true strain rate for test T17

48
 8 T18
x 10
9

6
true stress (Pa)

5 Max true stress 788 MPa

−1
−0.01 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09
true strain

0.15

0.1
true strainrate (/s)

0.05

−0.05
−0.01 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09
true strain

Figure 33 True stress and true strain rate for test T16

49
 8 T06
x 10
9

6
Max true stress 825 MPa
true stress (Pa)

−1
−0.01 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09
true strain

1.5

1
true strainrate (/s)

0.5

−0.5
−0.01 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09
true strain

Figure 34 True stress and true strain rate for test T06

50
 8 T07
x 10
9

6
Max true stress 856 MPa
true stress (Pa)

−1
−0.01 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09
true strain

1.5

1
true strainrate (/s)

0.5

−0.5
−0.01 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09
true strain

Figure 35 True stress and true strain rate for test T07

51
 8 T19
x 10
9

6
Max true stress 818 MPa
true stress (Pa)

−1
−0.01 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09
true strain

1.5

1
true strainrate (/s)

0.5

−0.5
−0.01 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09
true strain

Figure 36 True stress and true strain rate for test T19

52
 8 T20
x 10
9

6
true stress (Pa)

Max true stress 849 MPa

5

−1
−0.01 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09
true strain

1.5

1
true strainrate (/s)

0.5

−0.5
−0.01 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09
true strain

Figure 37 True stress and true strain rate for test T20

53
A2.2 Bolt load-extension curves
b06 (0.001/s filtered at 2 Hz)
100

90

80
max load 69.8 kN
70

proof load 61.0 kN

60 4.3 microns/kN
load (kN)

50

40

20−80% risetime 27 s
30 (12 − 49 kN)

20

10

−1 0 1 2 3 4 5 6
extension (mm)

Figure 38 Load extension curve for bolt test b06

54
 b07 (0.001/s filtered at 2 Hz)
100

90

80
max load 70.0 kN
70

proof load 61.9 kN

60 3.9 microns/kN
load (kN)

50

40

20−80% risetime 26 s
30 (12 − 50 kN)

20

10

−1 0 1 2 3 4 5 6
extension (mm)

Figure 39 Load extension curve for bolt test b07

55
 b15 (0.001/s filtered at 2 Hz)
100

90

80 max load 75.1 kN

70
proof load 64.5 kN
3.7 microns/kN
60
load (kN)

50

40

20−80% risetime 27 s
30 (13 − 52 kN)

20

10

−1 0 1 2 3 4 5 6
extension (mm)

Figure 40 Load extension curve for bolt test b15

56
 b08 (0.1/s filtered at 200 Hz)
100

90

80
max load 71.3 kN

70
proof load 63.4 kN
4.4 microns/kN
60
load (kN)

50

40

20−80% risetime 0.28 s

30 (13 − 51 kN)

20

10

−1 0 1 2 3 4 5 6
extension (mm)

Figure 41 Load extension curve for bolt test b08

57
 b09 (0.1/s filtered at 200 Hz)
100

90

80
max load 72.5 kN

70
proof load 62.9 kN
60 3.8 microns/kN
load (kN)

50

40

20−80% risetime 0.25 s

30 (13 − 50 kN)

20

10

−1 0 1 2 3 4 5 6
extension (mm)

Figure 42 Load extension curve for bolt test b09

58
 b10 (0.1/s filtered at 200 Hz)
100

90

80
max load 71.0 kN

70
proof load 64.2 kN
3.5 microns/kN
60
load (kN)

50

40

20−80% risetime 0.25 s

30 (13 − 51 kN)

20

10

−1 0 1 2 3 4 5 6
extension (mm)

Figure 43 Load extension curve for bolt test b10

59
 b11 (1/s filtered at 2000 Hz)
100

90

80
max load 72.9 kN

70
proof load 63.2 kN
3.7 microns/kN
60
load (kN)

50

40

20−80% risetime 0.027 s

30 (13 − 51 kN)

20

10

−1 0 1 2 3 4 5 6
extension (mm)

Figure 44 Load extension curve for bolt test b11

60
 b12 (1/s filtered at 2000 Hz)
100

90

80 max load 74.0 kN

70
proof load 63.0 kN
60 3.5 microns/kN
load (kN)

50

40

20−80% risetime 0.027 s

30 (13 − 50 kN)

20

10

−1 0 1 2 3 4 5 6
extension (mm)

Figure 45 Load extension curve for bolt test b12

61
 b13 (1/s filtered at 2000 Hz)
100

90

80
max load 73.1 kN

70
proof load 64.8 kN
3.3 microns/kN
60
load (kN)

50

40
20−80% risetime 0.027 s
30 (13 − 52 kN)

20

10

−1 0 1 2 3 4 5 6
extension (mm)

Figure 46 Load extension curve for bolt test b13

62
 b14 (1/s filtered at 2000 Hz)
100

90

80
max load 72.9 kN

70

proof load 61.2 kN

60 3.5 microns/kN
load (kN)

50

40

20−80% risetime 0.026 s

30 (12 − 49 kN)

20

10

−1 0 1 2 3 4 5 6
extension (mm)

Figure 47 Load extension curve for bolt test b14

63
A2.3 Shear tests

st2 (0.1/s filtered at 20 Hz)

100

90 max load 84.3 kN

80

70

60
load (kN)

50

40

30

20

10

−1 0 1 2 3 4 5 6
extension (mm)

Figure 48 Load-extension curve for shear test ST2

64
 st3 (0.1/s filtered at 20 Hz)
100

90 max load 83.9 kN

80

70

60
load (kN)

50

40

30

20

10

−1 0 1 2 3 4 5 6
extension (mm)

Figure 49 Load-extension curve for shear test ST3

65
 st4 (0.1/s filtered at 20 Hz)
100

90
max load 82.4 kN

80

70

60
load (kN)

50

40

30

20

10

−1 0 1 2 3 4 5 6
extension (mm)

Figure 50 Load-extension curve for shear test ST4

66
 st5 (0.1/s filtered at 20 Hz)
100

90
max load 81.0 kN

80

70

60
load (kN)

50

40

30

20

10

−1 0 1 2 3 4 5 6
extension (mm)

Figure 51 Load-extension curve for shear test ST5

67
 st6 (0.1/s filtered at 20 Hz)
100

90
max load 81.2 kN

80

70

60
load (kN)

50

40

30

20

10

−1 0 1 2 3 4 5 6
extension (mm)

Figure 52 Load-extension curve for shear test ST6

68