Materials Choices for Frangible Structural Elements

E. Brell* and C. Slattery**

*EW Brell & Associates
**Brisbane Airport Corporation

Abstract – Frangible structural elements intended to fail in the same direction as that of service load paths pose peculiar difficulties
in that the frangibility must be higher than the service loads. Ideally the frangible point would be a small margin of confidence above
the maximum service load in such situations. However, there are a number of factors in a bending frangible element that prevent
achievement of this ideal. The paper identifies what these factors are and discusses their influence on the relationship between service
limits and frangibility.

NOTATION
a Suffix for aluminium q Radius ratio s Suffix for steel
e Elongation R Outside radius U Ultimate strength
I Second moment of area Ri Inside radius UT Ultimate strength
MT Total moment of resistance Ro Outside radius Y Elastic limit
p Yield ratio (Y/U)

INTRODUCTION
Similar to most airports around the World, Brisbane Airport Corporation (BAC) has illuminated guidance
signs bordering runways. By their functional nature, they are required to be close to the runway edge,
making them prone to impact from errant aircraft. To minimize possible damage to aircraft, a mechanical
fuse or frangible element is interspersed in the support structure, to fail at a pre-determined force and pre-
determined height on the structure. However, the jet blast design wind velocities are stipulated up to 320
kph, demanding high service capacity from the frangible element. The directional service requirements are
the same as the normal flight path of aircraft. Where the directional requirement is at right angles to the
service load, frangibility may be less than the service load. Such a condition occurs in a roadside lamppost
where the cantilever load is at right angles to the flow of traffic. However, objects like road signs have
service loads (wind) in the same direction as impact load. In these types of structures the frangibility must
be higher than the service load. This can give rise to design frustration conditions where frangibility is so
high as to be valueless in mitigating impact damage.

By way of background, frangible structural elements on runway signs had been failing around the country in
response to a bow wave, rather than rear jet blast, of large aircraft, much to the surprise of arriving pilots
witnessing the collapse of signs ahead of their aircraft. The particular aircraft implicated in Brisbane were
the Boeing 747 and 777's. The Australian regulating authority, Civil Aviation Safety Authority (CASA)
gave dispensation to Sydney and Melbourne to relocate the signs further from the runway. Pre-emptively,
BAC viewed such concessions as short-term and chose to find long-term solutions instead. As a result,
frangibility was considered generally and choice of materials came into focus specifically. Guidance for the
design of frangibility for runway signs was originally provided by the International Civil Aviation
Organization (ICAO) in the Second Edition of Aerodrome Design Manual (1995) but was withdrawn
unheralded by the Third Edition (1999) as being in "preparation." The ICAO Frangible Aids Study Group,
after its last meeting in 1998, has not met again as the finalisation of a computer simulation model is awaited.
It is intended to offer this simulation technique as an alternative to impact tests for designing frangible
devices. [1].

There are a number of factors, which impinge on frangibility, emanating from material properties or
geometric peculiarities that maximize one material property to the detriment of the other. This paper
discusses some of these influencing factors. It is recognized that each such factor could be developed into a
full presentation study in its on right. The need for timely completion and brevity of presentation prevented
detailed treatise in some instances. The central objective of the paper is to give overview of the complex
factors that affect frangibility, the subject appearing sparsely covered in the literature. A typical
configuration of the BAC signs is shown in Figure 1.
 FRANGIBLE
COUPLINGS

Figure 1. Typical airport runway sign.

It can be observed from the sketch in Figure 1 that the frangible couplings exhibit a notch to localize failure
to a predetermined height. The notch is typically near the ground so that the sign falls, making a smaller
target for the remaining aircraft that may otherwise come into contact. The notch must be of sufficient cross-
section to resist the service bending load yet fail at the lowest possible impact force.

FRANGIBILITY DEFINED
The International Civil Aviation Authority [2] defines frangibility as follows:

The object is considered "frangible", if it breaks, deforms, or yields readily upon

impact and it is judged that the resulting damage to the impactor is such that no
hazardous condition exists.

For the purposes of the runway signs the definition is extended to ensure that the frangible couplings break
completely since a merely yielded sign is capable of damaging the aircraft further into the impact sequence.

OBJECTIVES
The objective of this paper is to assist in achieving maximum strength and maximum frangibility coping with
factors that impinge on these attributes. The chart in Figure 2 shows an ideal where the frangibility point is
at the yield point of the material.

It will be shown why the ostensibly brittle ideal in Figure 2 may not be achievable with engineering
materials that need post-processing such as machining. Some of the factors that may prevent achievement of
the ideal are as below:

1. Materials are stronger than specified by the manufacturer.

2. A large difference between yield strength and ultimate strength.
3. Notching geometry.
4. The role of resilience and toughness.
5. Dynamic response different to static response.

Ideal frangibility Realistic frangibility

Figure 2. Ideal frangible point.
MATERIALS SPECIFICATION
The metals supply paradigm is rooted in ensuring that the customer gets 'more' than is promised by published
values. This satisfies the great bulk of the marketplace, which is primarily interested in strength. This
principle is implicit in the term 'specified minimum yield' in common use by suppliers of metal.

The properties of the metal supplied can easily exceed the specified minimum yield and still be contractually
sound. However, in dealing with frangibility only the actual will satisfy as disclosed by coupon tests
supplied for the material batch, since the upper limit is as important as the lower limit of materials
specification. The final choice for Brisbane Airport was made from a batch with physical certificates. In
addition, a tensile coupon was pulled to verify the certificates. The result of the coupon test is posted below:
Table 1. Coupon test results EN26
Proof stress Ultimate tensile Elastic Elongation at Reduction
0.2% offset stress modulus failure of area
(MPa) (MPa) (Gpa) (% on 50mm) (%)
918 1024 196 20 59

Frangibility, by the above definition, occurs at the ultimate strength while strength for serviceability is
typically factored from the yield point (or 0.2% offset, as appropriate) of the material. A simple calculation
dividing the ultimate stress by the yield stress will show an ultimate/yield excess of 11.5% over what is
hoped for in the ideal in Figure 2. For the material tested in Table 1, frangibility would exceed the service
load by 11.5% plus an allowance for a confidence margin.

From the simplistic analysis above it follows that if the 11.5% excess was reduced, an immediate
improvement in frangibility ought to be achieved. A material that showed promise in this direction was a
solution-hardened aluminium with a 0.2% proof stress of 293 Mpa and an ultimate stress of 305 Mpa. A
similar simple calculation showed an ultimate/yield excess of only 4%. On the surface, this appeared a
superior material for the task at hand. It will be shown that this is not the case in the paragraph headed
"Material Comparison".

PREDICTION OF RUPTURE

In order to show that a material is superior or otherwise for frangibility purposes, the geometric context
needs to be defined. For the present application rupture of circular hollow section is chosen.

The behaviour of the section above the yield point becomes non-linear. It is not a simple matter to insert the
ultimate stress of the material into the classic bending equation (My/I) to predict the point of rupture. In
inelastic bending, assumptions must be made as to the distribution of the stress at rupture. Theoretical stress
distribution for a simple section is visualized in Figure 3.

A fully plastic hinge in a bending specimen exhibits a moment of resistance at failure equal to the ultimate
stress acting over half the area multiplied by the centroidal distance of the areas. [3]. Actual test specimens
(round hollow sections being considered) did not disclose any evidence of a fully plastic hinge at failure and
the idea of a fully plastic hinge was discarded. Instead a plastic circle segment held some appeal. This is
shown in Figure 5. Supporting this further was a finite element analysis, which showed a reasonably flat
distribution. It can be seen in Figure 4 that for a significant distance the isoplots (zones of single value) are
parallel (normal to centre line) to the section neutral axis.
 Figure 3. Simplified stress distributions ranging from elastic to fully plastic.

Figure 4. Bending stress distributions across section profile derived from finite
element analysis (redrawn to permit monotone presentation).

Figure 5. Assumed bending stress distributions across the full section profile showing
circle segments in plastic zone.

To test the validity of Figure 5 a coupling was proof tested in a compression-testing machine fitted with a
specially designed bracket to facilitate a cantilever load. The apparatus is shown in Figure 6. The theoretical
rupture load using the true stress from Figure 7 was calculated and compared with the actual recorded load
on the testing machine. Both these values are shown in Figure 8 where the calculated rupture is shown as a
line. Physical dimensions are tabled below:
Table 2. Physical dimensions of test piece
Outside diameter at notch root 47.0 mm
Inside diameter 40.0 mm
Moment arm 531 mm
Segment area 58.5 mm2
Segment centroid 21.4 mm
 Figure 6. Cantilever testing apparatus (hollow shaft end at fixture).
Gross shear stress was determined to be negligible. For simplicity, the "true" ultimate stress was calculated
from the coupon test using the final rupture force over the rupture area. No allowance was made for
tangential or radial stresses present in the necking section of the coupon. The coupon test result together
with the inferred "true" stress curve appears in Figure 7.

The calculation procedure adopted was to add the plastic moment of resistance to the elastic moment of
resistance of the remaining cross-section referred to as a double-dee section (see the shape of the double-dee
sections in Figure 11). The second moment of area of this section was calculated at 60,142 mm4. It was
assumed that the double-dee part of the cross-section was stressed to yield point at what would be the
extreme fibre were it not for the adjoining segments.

Part of the simplifying assumptions includes ignoring any shear that might arise between the segments and
the double-dee section. The sum of the elastic and plastic moments of resistance was calculated at 7,133
Nm, corresponding to a machine force of 13.4 kN at the 531 mm moment arm. This is shown in Figure 8.

Figure 7. Annotated coupon test result.

 16
14 CALCULATED RUPTURE
12
10

LOAD kN
8
6
4
2
0
0 7 14 21 28

DEFLECTION mm

Figure 8. Annotated bending test result.

Figure 9. Moment of resistance curves for circular hollow sections to give

overview of the equation and the general concept.

The calculated rupture point agrees with the test results to within 3%, suggesting that the plastic segment
approach may be valid. Work is continuing on a range of diameters and materials.

To facilitate the calculations, a single formula for total moment of resistance was developed using the basic
procedure as outlined previously. The formula is given without attempting to post the results of integration,
paucity of space precluding a detailed description of the mathematical formulations.

 ⌠ r⋅ 1 −q
2
 ⌠ r⋅ 1 −q
2

p ⋅U  π  4 4   ( ) 3   
(
 2 22 ) 2
⋅ ⋅ ( R) − ( q ⋅R)  − ⋅
4 2 2 3
M T :=
⌡
 R −x − ( q ⋅R)  dx  + 2 ⋅U ⋅
⌡0
 R − x − ( q ⋅R)  dx
q ⋅R  4 3 0  [1]

Where p is ratio of yield to ultimate strength, q is the ratio of inside radius to outside radius and R is the
outside radius. Equation [1] is plotted for yield to ultimate strength ratios from 0 to 1 in Figure 9. Yield
ratio p=0.5 is added to highlight the viability of linear interpolation between yield ratios.
MATERIAL COMPARISON
The purpose of this section is to highlight the geometric dependence of any comparison of materials for
frangibility service. The results of this study are to be compared with the simplistic ultimate/yield excess
study prior. Two cross-sections are compared, one in aluminium and the other in steel. The outside diameter
is held constant and the inside diameters adjusted to suit the yield strengths of the material. In this way the
yield moment of resistance is held constant for both sections representing identical service limits. The
material of cross-section with the lower rupture strength then is considered the superior material.

The unifying on yield for the two materials under consideration is accomplished by introducing fractions qs
and qa being decimal fractions of outside diameter in order to specify the inside diameter. Yield unification
is stated mathematically as follows:
Y a ⋅I a Y s ⋅I s
[2]

Substituting yield strength values and a literature expression for inertia of an annulus, equation [2] becomes:

π  4 π  4
⋅ ( R) − ( q a ⋅R)  ⋅ ( R) − ( q s ⋅R) 
4 4
293 ⋅ 918 ⋅
4 4 [3]

Eliminating R from the equation gives an expression of qa in terms of qs the radius fractions for aluminium
and steel respectively:

qa =
1 4
293
(
⋅ −15721098125 + 23091148926 ⋅qqs
4
)
[4]

The significance of equation [4] is that the radius fractions only and not the radii themselves, need to be
considered, to ensure that the extreme fibre is stressed to yield in each cross-section.

Equation [4] is plotted in Figure 10:

Figure 10. Radius fraction qa.plotted in terms of qs (both at respective yield).

Example: At yield point of aluminium (293 Mpa) the radius
fraction is say, at 0.75. At yield point of the steel (918 Mpa) for
the same loading the radius ratio only needs to be 0.94.
The analysis proceeds in five stages:

1. Calculate the elastic resistance moment to bring the relevant section to the material yield point
substituting equation [4] for the appropriate fraction.
2. Calculate the section resistance moment for the plastic sections.
3. Add the section resistance moments.
4. Compare total resistance moment with each other.
5. Compare the results with the simplistic ultimate/yield excess method.

Stage 1
To calculate the elastic resistance moment the classic (My/I) bending equation was used. The moment of
resistance was taken up to the yield point for each material. The results are shown in Figure 11

Stage 2
To calculate the section resistance moment of the plastic sections, the areas of circle segments were
multiplied by the ultimate stress for each material and the centroidal distance of the segment to the neutral
axis. Two segments contribute to the plastic moment of resistance (top & bottom). The results of such
calculations for internal diameters ranging from 0.91 to 0.99 of outside diameters for steel are plotted in
Figure 12

Stage 3-5
Figure 11 values are added to the values in Figure 12 and posted to the graphs in Figure 13.

Figure 11. Comparison of elastic moment of resistance for sections at respective yield points.

Figure 12. Comparison of plastic moment of resistance for sections at their respective ultimate stress.

Figure 13. Comparison of combined elastic and plastic moments of resistance for sections.
Figure 13 shows that for all values of internal diameter fractions from 0.91 to 0.99 (steel) and corresponding
fractions for aluminium as determined by ratios previously discussed, the steel samples rupture sooner than
the aluminium samples, yet have the same moments of resistance to yield. The particular grade steel, being
closer to the ideal in Figure 2 is the superior material over the particular grade aluminium, where, depending
on the radius fraction, the improvement is up to 28%. In contrast, the simplistic non-geometric study showed
an improvement of the aluminium over the steel.

STRESS CONCENTRATION
The literature abounds with studies that compare a notched specimen with an equal cross-section but
unnotched. The stress concentration factors arising from such comparisons apply to the elastic condition.
During plastic working, the very local stress intensity is relieved to the flow stress of the material. Some
form of notching is unavoidable on a frangible coupling since localization of fracture is necessary to give a
degree of determinacy to the frangibility process. Since stress concentration reduces the yielding force, it is
an undesirable by-product of such fracture localization requirement.

If sharp notching is unavoidable, some caution is needed in applying these stress concentration factors. The
derivation of these factors may be in context of fatigue prediction where a minimum specification paradigm
may exist that would permit under-prediction with immunity, analogous to the materials supply paradigm.

A trade-off between elastic stress concentrator and the effects of material toughness is discussed below.

RESILIENCE AND TOUGHNESS

The ability of a material to absorb energy under adverse conditions is a complex function of its basic
mechanical properties (stress-strain relations, including time effects.) Because it is difficult to predict such
behaviour analytically, special tests have been devised which measure the energy absorbed in impact
loading. The results of such tests are not, in themselves, material properties, but they are useful in selecting
materials, in making direct comparisons between different materials. [4]. More recently, advances made in
predicting material toughness from instrumented Charpy tests have extended the usefulness, but are
restricted to families of materials. There appears no universally applicable correlation formula to date [5]
and [6]. The Automatic Ball Indentor (ABI) is able to determine toughness non-destructively to a reasonable
and verifiable accuracy. [7] and [8]

To discuss resilience and toughness, the concept of participation length is worth reviewing. In a tensile test
coupon, the participation length is eminently visible even though it is visually misrepresented by the extent
of elongation of the coupon. This is illustrated below:

The participation length is the distance between the curvature tangents where they meet the plastically
undeformed diameter of the specimen less the elongation of the specimen. It is shown in Figure 14 but
pictorially and dimensionally includes the elongation (e) to rupture.

The participation length of a grooved hollow frangible coupling is more difficult to determine. To discuss
the effect of notches in the present context, it is convenient to take an elemental strip, which contains the
extreme fibre. The strip notion completely ignores shear connection to its adjoining strip, the notion's virtue
being in communications rather than mathematical rigour. Also, such a tensile strip, if pulled in isolation,
would exhibit bending. The notional strip is unaffected by any bending propensity.
 Figure 14. Tensile test coupon sketch and photo showing participation length including elongation - e.

Figure 15. Coupling showing elemental strip with participation length and photo of actual coupling.

Modulus of Resilience
Texts on engineering materials will show the modulus of resilience as a measurement of energy in terms of
volume (Joules/mm3). It is determined by calculating the area under the stress-strain curve bounded by the
origin and the elastic limit of the material multiplied by the elastic participation volume. It is a measure of
the amount of strain energy a specimen must absorb before the material will yield. The area under the elastic
section of curve is small compared to the total area under the curve. An appreciation of proportion can be
seen in Figure 7 where the total area under the curve is approximately 20 times the resilience.

Wind, whether weather or airplane generated, comes in gusts. A resilient design is more able to withstand
these gusts without giving rise to a need to increase the service design load. Resilience in a material is seen
therefore as a beneficial attribute.

Modulus of Toughness
Toughness is also a volumetric measure, taken over the plastic participation volume. To illustrate how
toughness affects frangibility, two notional strips of finite thickness are considered as discussed prior. Each
strip is notched with a large or a small radius to contra-distinguish by a wide margin. This is shown in
Figure 16 where a non-participation thickness is chosen beyond which plastic strain is not taking place. The
dark shaded area represents the plastic working volume over which the material's modulus of toughness is
applied. It is immediately apparent that the strip with the large radius is clearly able to absorb more plastic
strain energy prior to rupture than the adjoining strip with a much smaller area.

Figure 16. Toughness concept applied to notch root radius.

Toughness effects occur above the yield point (where fracture would have already occurred in the ideal
specimen of Figure 2). Because toughness delays fracture it is an undesirable material quality for frangible
couplings. The choice of root radius affects the overall toughness of the coupling– the more participation
volume, the greater the delay in the arrival of fracture, the greater the potential for aircraft damage. This
exposes an important trade-off in the choice of root radius of the notch. A large radius is undesirable from a
toughness perspective while a small radius is undesirable from a stress concentration perspective.

DYNAMIC MATERIAL PROPERTIES

Mechanical Impedance
Mechanical impedance accounts for the inertial response in the interval of time between impact and free-
body motion. The effect of impedance can be considered in analogy, as people standing in a line where one
person at one end gets "crushed" by a moving vehicle. It takes a finite time to get the message to the other
end of the line to get the last person to move. Then there is time needed for each person to transfer into the
space made by the end person to the adjoining person and finally, to the person in need of the space. The
faster the vehicular speed, the greater the "crush" of the first person. The idea is illustrated in caricature in
Figure 17.

Stress of a specimen is proportional to velocity of impact with the constant of proportionality being
impedance. Impedance is the product of material density and celerity, often referred to a material's acoustic
velocity. Celerity relates to a strain (elastic) disturbance, which in aluminium and steel progresses at
nominally 5 km/s. However, since the density of aluminium is approximately one third of steel density, the
localized stress is less in aluminium as compared to steel.

An elastic impedance manifestation is unlikely to have any effect on the serviceability performance of the
coupling. However, plastic impedance is nominally 10% of the elastic impedance [9] and may manifest as
critical velocity.

Critical Velocity.
It has been known since the late 40's and early 50's that longitudinal plastic deformation is trapped by
longitudinal plastic waves giving rise to critical impact velocity. [10]

There exists a velocity at which a tensile specimen will always fail, under impact, independent of cross
section of the specimen or the mass of the impacting mass, provided that the kinetic energy is at least
sufficient to supply the energy absorbed by the necking-down process associated with tensile failures. [11]

Figure 17 Impedance analogy.

To get a quantitative appreciation, some critical velocities are posted in Table 3.
Table 3. Critical velocities of some materials.
ULTIMATE STRENGTH ELONGATION
MATERIAL CRITICAL VELOCITY
STATIC DYNAMIC STATIC DYNAMIC
SAE1045 Q&T 985 MPa 1165 MPa 5.7% 9.2% 80m/s
2S ALUM ½ H 119 MPa 152 Mpa 4.6% 7.0% 33m/s
Units converted from (Clark 1953)

These velocities are generally high and are thought to be out of range of frangible couplings.

Strain Rate
Strain rate effects are characteristically responsible for higher flow stresses. This is shown schematically in
Figure 18.

To determine whether the airport signs were in the general domain of such effects, line sketches were drawn
for signs subjected to wind loading conditions and also impact from an aircraft nose wheel, the latter due to
[12]. Mental arithmetic was employed to determine crude coupling notch separation velocities. These were
determined to be approximately 4.5 m/s for the wind load condition and 0.5 m/s for the nose wheel impact
condition, as shown below:

If the participation length were only 1 mm, strain rates of 4500/s and 500/s respectively for the wind loading
and nose wheel impact, are feasible. Such strain rates may be high enough to be within range of dynamic
responses of some materials for frangible couplings. Under conditions of wind loading a momentary peak as
shown in Figure 18 may be desirable. Under impact, strengthening delays rupture and is therefore
undesirable. Accordingly, material sensitivity to rate of strain may need to be considered either in the choice
of materials or in the determination of participation length through root radius specification.

50,000/s

10,000/s
STRESS

5000/s
500/s

STRAIN
Figure 18. Schematic evolution of stress-strain curves of
industrial steels at high strain rates [10].

Figure 19. Line diagrams of sign arrangements

for a 50mm diameter coupling.
CONCLUSIONS
Throughout the paper it was shown how to manipulate the factors that bring the frangibility as close as
possible to the yield point of the material. It was shown that a mere comparison of ultimate strength to yield
strength to be misleading with regard to frangibility, geometry having emphasized one material attribute to
the detriment of another. It was thus shown that geometry could be a factor in materials selection to achieve
optimum frangibility.

The role of notching was discussed in context of toughness and stress concentration, concluding that a large
notch radius improved serviceability but increased frangibility. Conversely, it was found that a small notch
radius improved frangibility but decreased serviceability. Scope for notch radius optimisation was shown to
be warranted. Elastic and plastic impedance were found to be relatively unimportant in the design of
frangibility in the general configuration of Figure 1.

Strain rate effects may be a factor in the choice of some materials. The above notch radius optimisation may
have implications in mitigating the effects of high strain rate manifestations.

A modulus of rupture showing good test agreement is offered as a seed for future study. Curves based on
this modulus of rupture for frangible hollow circular section are presented however, there being only one
corroborating test, the curves are subject to further verification.

ACKNOWLEDGEMENTS
Dr Martin Veidt, Department of Mechanical Engineering, University of Queensland.
Mr Barry Neville, Department of Civil Engineering, University of Queensland.
Mr Colin Wolf, Techniplan Pty Ltd

REFERENCES
1. A.K.R. Rao Email to E. Brell, 2001, Chief, AGA Section, ICAO: Montreal. Canada.
2. ICAO-Ed2, I.C.A.O., Aerodrome Design and Operations. 1995.
3. W.C. Young Roark's Formulas for Stress and Strain. 6th Edition ed. 1989: McGraw Hill.
4. F.R. Shanley Strength of Materials. 1957: McGraw-Hill.
5. H.J. Schindler and U. Morf, A Closer Look at Estimation of Fracture Toughness from Charpy V-
Notch Tests. Int J Pres Ves & Piping, 1993: p. 203-212.
6. H.J. Schindler and M. Veidt, Fracture Toughness Evaluation From Instrumented Sub-Size Charpy-
Type Tests. ASTM STP 1329, W.R. Corwin, S.T. Rosinski, and E. van Walle, Eds. American
Society for Testing and Materials., 1997.
7. F.M. Haggag How can fracture toughness of ferritic steels be determined from the ABI test?. 2001,
Advanced Technology Corporation, Inc.
8. T. Byun, et. al, A Theoretical Model for Determination of Fracture Toughness of Reactor Pressure
Vessel Steels in the Transition Region from Automated Ball Indentation Test.
Journal of Nuclear Materials, 1998. 252: p. 187-194.
9. W. Johnson Impact Strength of Materials. 1972, London: Edward Arnold.
10. J. Klepaczko Plastic Shearing at Very High Strain Rates, A Review. in IUTAM Symposium. 1995.
Noda Japan: Springer.
11. H.A. Rothbart Mechanical Design Handbook. 1996: McGraw-Hill.
12. G. Eriksson Test Report Horizontal Impact Tests on a Taxiing Guidance Sign. 1995,
Swedish Board of Civil Aviation.