Lightweight Wine Bottles

Transcription

Final Report: GlassRite: Wine
Strength as an Issue in the Manufacture of Lightweight
Wine Bottles
Project code: MSG009
Research date: July 2007 to February 2008
ISBN: 1-84405-388-1
Date: May 2008
WRAP helps individuals, businesses and
local authorities to reduce waste and
recycle more, making better use of
resources and helping to tackle climate
change.
Written by: Andy Hartley
Front cover photography: View of standard and lightweight wine bottles.
WRAP and GTS believe the content of this report to be correct as at the date of writing. However, factors such as prices, levels of recycled content and regulatory
requirements are subject to change and users of the report should check with their suppliers to confirm the current situation. In addition, care should be taken in using
any of the cost information provided as it is based upon numerous project-specific assumptions (such as scale, location, tender context, etc.).
The report does not claim to be exhaustive, nor does it claim to cover all relevant products and specifications available on the market. While steps have been taken to
ensure accuracy, WRAP cannot accept responsibility or be held liable to any person for any loss or damage arising out of or in connection with this information being
inaccurate, incomplete or misleading. It is the responsibility of the potential user of a material or product to consult with the supplier or manufacturer and ascertain
whether a particular product will satisfy their specific requirements. The listing or featuring of a particular product or company does not constitute an endorsement by
WRAP and WRAP cannot guarantee the performance of individual products or materials. This material is copyrighted. It may be reproduced free of charge subject to the
material being accurate and not used in a misleading context. The source of the material must be identified and the copyright status acknowledged. This material must
not be used to endorse or used to suggest WRAP’s endorsement of a commercial product or service. For more detail, please refer to WRAP’s Terms & Conditions on its
web site: www.wrap.org.uk
Executive Summary
This report has been prepared as part of the WRAP-funded GlassRite suite of projects which are intended to
reduce the volume of packaging entering the UK waste stream. The GlassRite projects aim to encourage the
lightweighting of glass containers as a means of reducing the tonnage of glass in the waste stream. The
GlassRite Wine project has the additional objectives of seeking to promote the bulk importation of wine into the
UK to be filled in lightweighted bottles and also of encouraging importers of bottled wine to consider the use of
clear glass.
This report seeks to demonstrate that lighter, thinner wine bottles can now be manufactured to a strength equal
to that of the heavier bottles that are in common usage. The report also seeks to demonstrate that computational
aids such as Finite Element Analysis (FEA) can usefully be employed to evaluate new designs prior to their
manufacture and identify any potential structural weaknesses.
The report provides potential users of glass with an understanding of the factors that influence glass strength. It
gives an insight into the improved manufacturing techniques that have allows glass manufacturers to reduce the
weight of containers without compromising strength or safety.
Glass has many properties that make it an ideal choice for wine bottles and for packaging many other foodstuffs
but, whilst it is an inherently strong material, it is also a brittle one. The report seeks to explain how and why a
brittle material like glass fails.
Glass containers are produced by blowing the glass into metal moulds. The amount of glass used to make the
container is a secondary consideration in determining the final strength. The principal consideration is to ensure
that the glass that has been used is properly distributed and that no thin spots occur. Preventing manufacturing
flaws and minimising surface damage are also important factors that determine a container’s strength.
As a means of demonstrating that lighter wine bottles can be manufactured to a comparable strength to their
heavier counterparts, a comparison has been made between three 750ml Bordeaux style bottles at weights
ranging from 360 to 475g. The comparison has been made using FEA with some physical testing used to validate
the process. The results of the analysis demonstrate the usefulness of computational tools in the design process
and show that the lighter bottle is perfectly serviceable.
Reference is made in the report to some other practical research in which a more direct comparison was made
between a typical container and its lightweighted counterpart. The work concluded that, in general, the lighter
container was actually the stronger container.
The report also demonstrates a wireless sensor technology which offers the opportunity to test the performance
of containers in real time in real environments e.g. on the filling line. The technology is not yet in general use but
could become an invaluable design tool leading to better understanding of the products lifecycle and thereby help
designers make further weight reductions.
1
Contents
1.0
2.0
Introduction ............................................................................................................................. 2
Why do Glass Containers Fail? ................................................................................................. 3
2.1
Glass Failure .........................................................................................................................3
2.2
Glass Flaws and Surface Damage ...........................................................................................3
2.3
Stress Inducers .....................................................................................................................5
2.4
Glass Failure Modes ...............................................................................................................5
3.0
The Safe Manufacture of Lightweight Glassware .................................................................... 7
3.1
Improved Design Capabilities .................................................................................................7
3.2
Improved manufacturing process control ................................................................................8
3.3
Improved Mould Production ...................................................................................................8
3.4
Improved Machine Operation .................................................................................................8
3.5
Improved Inspection Technology............................................................................................8
3.6
Real Time In-situ Monitoring ..................................................................................................8
4.0
The Performance of Lightweight Wine Bottles ........................................................................ 9
4.1
Finite Element Analysis ..........................................................................................................9
4.2
FEA Comparison of Three Wine Bottles of Different Weights .................................................. 10
4.2.1
Inputs .................................................................................................................... 10
4.2.2
Outputs.................................................................................................................. 10
4.3
Results ............................................................................................................................... 11
4.3.1
Internal Pressure and Head load .............................................................................. 11
4.3.2
Impact Loading....................................................................................................... 12
4.3.3
Physical Testing ...................................................................................................... 16
5.0
Other Comparative Studies of Lightweighted Container Strengths....................................... 16
6.0
Conclusions ............................................................................................................................ 17
Appendix A: Modelling Results........................................................................................................... 18
Appendix B: Physical Testing Data & Results..................................................................................... 24
Figures
Figure 1 The Blow and Blow glass forming process. ......................................................................................... 4
Figure 2 The Press and Blow glass forming process. ........................................................................................ 4
Figure 3 Residual Stress in a glass container. .................................................................................................. 5
Figure 4 Tensile stresses produced in hollow items by squeezing...................................................................... 6
Figure 5 External impact (a) and internal pressure failure (b). .......................................................................... 7
Figure 6 “Wireless” monitoring of a filling line using a mock bottle with series of sensors. ..................................9
Figure 7 Combined pressure and head load for wine bottle A. ........................................................................ 11
Figure 8 Impact simulation conditions. .......................................................................................................... 13
Figure 9 Impact loading for wine bottle A...................................................................................................... 13
Figure 10 Comparison of physical testing and finite element analysis (shoulder strikes only). ........................... 15
Figure 11 Glass Distribution of the three wine bottles investigated................................................................... 18
Figure 12 Combined pressure and head load for container A (lightest weight). ................................................ 19
Figure 13 Combined pressure and head load for container B (intermediate weight). ........................................ 20
Figure 14 Combined pressure and head load for container C (heaviest weight)................................................ 20
Figure 15 Impact loading for container A....................................................................................................... 21
Figure 16 Impact loading for container B ..................................................................................................... 22
Figure 17 Impact loading for container C. ..................................................................................................... 23
Tables
Table 1
Table 2
Table 3
Table 4
Physical dimensions of wine bottles investigated................................................................................ 10
Maximum stresses from combined pressure and head loading. ........................................................... 12
Maximum tensile stresses for impact stresses for the three wine bottles investigated........................... 15
Summary of the physical impact testing. ........................................................................................... 16
1
1.0
Introduction
This report seeks to provide those involved in the purchase and specification of wine bottles with information
relating the strength of glass bottles. It aims to demonstrate that modern glass manufacturing methods can
produce lighter, thinner wine bottles which have a comparable strength to those heavier versions that have
traditionally been used by the wine trade. Wine bottles are a major contributor to the UK’s waste stream and
WRAP’s Glassrite suite of projects aims to reduce this impact. Reducing the weight of packaging is one means of
achieving this objective and this report seeks to promote lightweighting activities by providing clear evidence that
lightweight bottles are safe to use and not more prone to failure than their heavier counterparts.
To be commercially viable, glass containers must be produced at high speeds by automated machines. Forming a
hollow item such as a glass container is inherently more complicated than making a solid product. The glass
forming process essentially entails introducing a relatively small amount of molten glass into a mould and then
blowing the glass outwards to fill the mould space. Calculating how much glass to place into the mould is not
difficult; the challenging task is to ensure that glass is distributed appropriately within the mould and that no
unwanted thin spots occur which create areas of weakness.
Glass is a strong material but due to a lack of crystalline structure it is brittle. The container manufacturer will be
aware of the conditions and environment to which the container will be exposed and the stresses that it is
expected to withstand. The manufacturer will also be aware that the glass forming process will inevitably have
resulted in microscopic surface damage and may have introduced other flaws into the body of the glass.
One of the most challenging aspects of glass container production is ensuring
that the glass is properly distributed in the mould. In the absence of good
control over glass distribution the glassmaker has to resort to making thickerwall containers, in the hope that even the thin spots have sufficient glass to
ensure that the container will not fail in use.
The development of the Narrow Neck Press and Blow (NNPB) manufacturing
process allowed glassmakers to produce lighter, stronger bottles as it greatly
increased the control they had over glass distribution. The process was initially
used for the high volume beer market and resulted in dramatic weight
reductions. By adopting the new technology, companies were easily able to
achieve weight reductions of 20% with no loss in strength or any need for a
radical change to the container’s shape.
As an example, the Ukrainian-based Gostomel Glass Factory (GGF) adopted the
technology and was able to report “The new production lines will allow GGF to
decrease the weight of beer bottles from the current range of 340g to 380g to
ultra lightweight bottles weighing 280g to 240g (a substantial saving in raw
materials) while retaining the 15 bar pressure rating.”.
Adnams beer bottle
Spirit bottles have also been lightweighted. Until recently the tall round 70cl industry standard was being
produced at around 450g. The same bottle is now available for general use at 298g. The Co-op1 was able to
claim to be the first company to adopt the new “sub-300g environmentally friendlier” bottle by using it for their
own-brand whisky.
Having gained control of glass distribution the container designers are now better placed to benefit from
computer modelling technology.
Finite Element Analysis (FEA) is one of the most widely used modelling tools finding applications in many areas of
advanced engineering. FEA enables a designer to test a new container for a range of performance indices such as
vertical loading, bursting pressure and impact resistance without the need for expensive prototypes.
1
Spirit of Innovation: Co-op Roll Out World's Lightest Whisky Bottle,
http://www.wrap.org.uk/wrap_corporate/news/spirit_of.html
2
Coca-Cola is perhaps the most high profile company to acknowledge the value of FEA in container design. The
Coke bottle is an iconic shape and the marketing implications of any change to the profile would be profound.
However, the Coke bottle has been redesigned with the aid of FEA. The new bottle is seen by Coca-Cola as part
of their commitment to corporate social responsibility.
The Ultra Glass contour bottle is designed to improve impact resistance, and reduce weight and cost. The innovative Ultra
Glass bottles are 40 percent stronger, 20 percent lighter and 10 percent less expensive than traditional contour bottles.
Use of the Ultra Glass design has eliminated 52,000 metric tons of glass – resulting in a CO2 reduction of 26,000 tons or
the equivalent of planting 8,000 acres of trees.2
2.0
Why do Glass Containers Fail?
Glass technologists frequently claim that their material is “inherently strong”, yet common experience teaches us
that the material is fragile. How can the claim of strength be reconciled with contradictory the experience of
fragility?
An answer to this quandary is provided by gaining a better understanding of the mechanism of glass failure. A
better understanding of glass properties will also help glass users identify possible problems in handling and in
specifying glass containers.
2.1
Glass Failure
Glass containers fail when sufficient stress is applied. The glass invariably fails by cracking i.e. splitting apart.
Glass in pristine condition will be able to resist large stresses. Unfortunately, any flaws will significantly reduce the
ability of a glass container to withstand stress.
However, the application of stress to a glass flaw does not necessarily equate to failure. It is only those stresses
that are acting to pull the glass apart (tensile stresses) and exacerbating flaws and tiny cracks that will cause
failure. Stresses (compressive) that act to close a flaw such as surface crack will not cause failure.
The failure of a glass item could be described by the following simple expression:
Glass Flaw + (High) Tensile Stress = Failure
2.2
Glass Flaws and Surface Damage
The strength of a glass bottle can be reduced by flaws generated during its manufacture and by surface damage
sustained in subsequent use. Flaws can be introduced into glass bottles during manufacturing, some of which
are serious and can compromise the strength of the finished product. Some of these flaws are due to poor
process control e.g. thin walls; others are due to contaminants being introduced into the furnace such as
inclusions from broken ceramic often found in the cullet supply and which can pass through the manufacturing
and end up embedded in a bottle’s sidewall or base. Fortunately, the rigorous inspection process which every
bottle undergoes is able to detect the great majority of these flaws before the bottle leaves the glass factory.
Bottles are formed in metal moulds. Two basic forming processes are used to produce bottles; the “blow and
blow” and the “press and blow” methods. Both processes employ two moulds to form an article; the blank mould
and the blow mould. The blow and blow process is shown in Figure and essentially uses a small metal plug to
begin the process and form the neck and relies solely on air pressure to form the bottle. The press and blow
process is shown in Figure and involves pressing a metal plunger into the glass prior to the start to the “blowing”
stage. The press and blow process is more complex but gives better control over glass distribution.
Both processes involve glass to metal contact which results in the bottle suffering a small amount of surface
damage and whilst under normal conditions the bottle will be perfectly safe and serviceable, it will have lost some
of its strength.
2
Package Design, http://www.thecoca-colacompany.com/citizenship/package_design.html
3
Figure 1 The Blow and Blow glass forming process.
Gob
dropped
into blank
mould
Neck
formed
Blank
blown
Blank
shape
Blank
transferred
to blow
mould
Final shape
blown
Finished
bottle
Figure 2 The Press and Blow glass forming process.
Gob
dropped
into blank
mould
Plunger
presses
blank
shape
Blank
pressed
Blank
shape
Blank
transferred
to blow
mould
Final shape
blown
Finished
jar
Once formed the bottles passes through an annealing process which involves controlled heating and cooling
stages and which is intended to remove the internal stress generated during forming process. Ideally annealing
the glass will remove all the internal stress but in practice some will remain. Thicker glass needs to spend more
time in the annealing oven and the annealing time is thus determined by the maximum wall thickness of the
container. Lightweighted containers, being thinner and in general having a more uniform glass distribution than
their heavier counterparts, require less annealing and are thus less prone to being poorly annealed.
Consequently, glass manufacturers include tests for annealing in their quality control regime of tests.
Fortunately, any residual stress is easily identified by simply viewing the container under polarized light and any
stress present can be quantified by comparison to standard strain discs. Figure shows a jar that has some
residual stress in the shoulder and base (the stressed areas appearing yellow/orange under polarized light).
4
Figure 3 Residual Stress in a glass container.
Once annealed the bottles progress through several inspection stations before being palletised for despatch to
the customer. All of these stages involve some form of glass to glass or glass to metal contact which adds to the
surface damage.
Whilst glass is not unique in that it suffers from surface flaws, it is unusual in having a non-crystalline structure,
which makes it particularly susceptible to tensile failure. Other materials are essentially composed of many
interlocked grains which present a barrier to crack propagation. Glass lacks this internal grain structure and thus
allows cracks to develop unhindered.
2.3
Stress Inducers
In normal use, glass containers will be subject to a variety of stresses which could act on a glass flaw to produce
breakage. These stresses include:
2.4
Impacts
bumps and knocks from filling and general use
Squeezing
sideways pressure from rollers on filling lines
Loading
vertical pressure from stacking
Residual stress
internal manufacturing stresses that should be removed by annealing
Bursting
internal pressure from carbonated drinks - which is of particularly relevance to
champagne and sparkling wine production that involves re-fermentation
Thermal
rapid expansion due to hot filling
Glass Failure Modes
Glass fails by cracking and forces that act to close a split (compressive) will not cause failure but those that act to
pull apart the split (tensile) may cause failure.
Forces acting on hollow items such as containers can simultaneously produce both compressive and tensile
stresses. A force that produces compressive stress on the outer surface also produces tensile stress on the inner.
Figure shows a container that is being placed under an external compressive load; it is being squeezed as might
occur between the rollers on a filling line. The squeezing action is tending to compress the container into an
elliptical shape. The outer surface at the compression point is being flattened whilst the inner surface is being
stretched i.e. is being put under a tensile stress. At 90 degrees to the applied load the converse situation applies.
Here the outer surface is being stretched whilst the inner surface is being compressed. Points were the glass is
under high tensile stress and thus tending to being pulled apart are known a “hinge points” and are the positions
that are most prone to failure.
5
Thus, as the outer surface of a container is more likely to contain a flaw than the inner, a container that is
crushed may fail at a point 90 degrees to the application of the pressure.
Figure 4 Tensile stresses produced in hollow items by squeezing.
Inner surface being
Outer surface being
compressed
stretched (tensile)
Hinge Point
Squeeze
Squeeze
Inner surface being
compressed
Outer surface being
stretched (tensile)
Hinge Point
When glass fails it leaves behind characteristic features on the fracture surfaces which can provide clear evidence
as to origin and cause of failure. Trained observers can reconstruct items and, with their specialist knowledge of
crack propagation, can trace failure patterns back to the point of origin. The origin of a glass fracture may not
coincide with the point at which the container was struck or stressed; however, the experienced observer can
usually determine the actual cause of a failure which may be a combination of a glass fault and an applied stress.
With the evidence provided by the fragments an investigator can often determine if a failure was due to a design
fault or the misuse or abuse of the failed item.
Figure shows two examples of failure, one being a typical external impact, the other the result of internal
pressure as might occur with a sparkling wine bottle. More examples of typical failure patterns can be found in
the publication “The American Society for Testing and Materials (ASTM) Designation: C 1256 – 93 “Standard
Practice for Interpreting Glass Fracture Surface Features”.
6
Figure 5 External impact (a) and internal pressure failure (b).
Impact Failure
Star pattern
a) outer surface (by impact)
3.0
Pressure Failure
Forked pattern
a) inner surface (by pressure)
The Safe Manufacture of Lightweight Glassware
Given that glass is a brittle material and can fail when it suffers a sharp blow or is simply dropped, is
lightweighting a prudent step?
Glassmakers are confident in their ability to reduce the weight of containers whilst retaining (and in some cases
improving) structural strength. The “lightweighting process” is viewed as the natural outcome to the progressive
application of new and improved technology to the manufacturing process. In recent times significant advances
have been made in better design, forming and inspection and testing techniques all of which have all contributed
to the safe production of lighter glassware. A brief description of these improvements is given in the following
sections.
3.1
Improved Design Capabilities
The contribution to the lightweighting process of computer-assisted design (CAD) tools has been significant. Glass
container designers were some of the first users of CAD technology. Initially the use of these programmes was
confined to producing images for customers and more detailed technical information to assist with mould
production.
Further increases in computer power and the ready availability of user-friendly packages eventually made the use
of Finite Element Analysis (FEA) a practical reality for glass container designers. Several software packages are
now commercially available. Some consider the actual forming process and how the glass moves during forming,
which helps the designer direct more glass into potential weak spots; others consider the performance of the
finished container. FEA enables the designer to model the strength implications of any design feature and reduce
the need for the costly step of actually producing prototype glass containers for testing and subsequent bottle
redesign that may be necessary.
7
3.2
Improved manufacturing process control
Glassmakers can never achieve perfect glass distribution within a mould and must incorporate a safety factor to
ensure that glass thickness does not fall below a safe level at any point. The safety margin is needed because of
the uncertainty in ensuring good distribution. Today’s glassmakers have much more control over the forming
process. The glass forming process begins when a discrete portion of molten glass (a gob) is fed to the mould. A
consistent gob weight is a prerequisite for a uniform product. Advances in the control of the parameters which
determine gob weight have reduced the variations in the amount of glass that is fed to the mould.
Better control is also exercised over the movement of the plunger used in the Narrow Neck Press and Blow
(NNPB) process which results in better glass distribution within the mould.
The improved level of control at these two critical points gives the glassmaker confidence that the a safe glass
distribution has been achieved with the result that a lower safety margin is required and less glass need be used.
3.3
Improved Mould Production
The cast iron moulds used for forming the glass containers are negatives of the required shape with allowances
for thermal expansion at the glass forming temperature. Improved machining technology to manufacture the
moulds now enables the moulds to be produced to much tighter tolerances.
3.4
Improved Machine Operation
The glass forming machine performs a complex sequence of actions in order to produce a wine bottle or jar (see
Figure 1 The Blow and Blow glass forming process.
Modern forming machines use electronic timing mechanisms which give more precise control over the timing of
individual events than was possible with old mechanical timing technology. This improvement in process control
again helps the manufacturer to work to tighter tolerances and a lower safety margin and thus allowing the use
of less glass.
The cast iron moulds used to shape the containers need continuous and rapid cooling. Best practice now involves
controlled internal mould cooling which gives much better control than the older external blow cooling method.
3.5
Improved Inspection Technology
Every container produced is subject to a range of quality checks during the manufacturing process. Increasingly
these tests are performed in-situ, by technology based on high speed imagery which is able to identify and reject
any manufacturing faults including: thin walls, inclusions, crizzles (small cracks) and dimensional variations. A
number of containers are also routinely tested to destruction to assess impact resistance, vertical load and
bursting pressure the latter being of particular importance to highly carbonated sparkling wines.
Thus the quality system ensures that any lightweighted container meets the same exacting standards that were
applied to its heavier predecessor.
3.6
Real Time In-situ Monitoring
Developments in telecommunications and data logging now enable the glass manufacturer to gather information
about the stresses that the glass container will be subjected to during its working life. One such system involves
producing an exact perspex replica of the container within which are installed various load, temperature, and
accelerometer impact sensors. The model container is then introduced into the production line and is able to send
back information in real time on the conditions that prevail (Figure ). The information gathered can be used to
refine the design of the container or alternatively can identify those points in the production, filling and
distribution system at which the container is at risk. To date three UK glass manufacturers have acquired this
technology but no operational details have yet been made available.
8
Figure 6 “Wireless” monitoring of a filling line using a mock bottle with series of sensors.
The data received from the sensing equipment relating to impacts is not directly comparable with the standard
pendulum impact testers found in quality laboratories, but it is arguably more representative of actual operating
conditions. The laboratory equipment places a container against an immovable V-backstop and subjects it to a
standard blow. In reality a container will receive only glancing blows from fixed objects and bump into other
(moveable) containers on the filling line. It is also important to recognise that for a moving container the energy
of the impact is proportional to the container weight so lighter containers will receive smaller knocks causing less
damage.
4.0
The Performance of Lightweight Wine Bottles
As a means of demonstrating that lighter wine bottles can be manufactured to a comparable strength to their
heavier counterparts, the performance of 3 existing wine bottles has been evaluated. The bottles chosen had
weights ranging from 360 to 475g and the analysis was designed to determine their relative ability to withstand
the stresses that such bottles could reasonably be expected to encounter in normal use. The bottles selected
were three 75cl Bordeaux style wine bottles of varying weight and glass thickness.
The performance of the bottles was compared by mathematical modelling using a standard Finite Element
Analysis (FEA) package. The bottles were also subjected to standard impact testing as would be applied in the
glass factory as a means of validating the FEA output.
4.1
Finite Element Analysis
Finite element analysis is a computational method that is increasingly finding application in glass container
design. The technique makes it possible to evaluate strength and performance of a glass container on a
computer before final design decisions are made. In the absence of FEA (or other numerical analysis), decisions
on safe wall thicknesses and the radii of curved surfaces were based on experience supplemented by hand
calculations. For a complex structure, the simplifying assumptions required to make any calculations possible
often led to a conservative and over heavy design. A considerable factor of ignorance could remain as to whether
the design would be adequate for all expected loads.
9
4.2
FEA Comparison of Three Wine Bottles of Different Weights
The FEA study was undertaken by Rockfield Software which has recognised expertise in the field of FEA and
which is used by some UK glass manufacturers in their development work. In order to perform a FEA the
following details were required:
4.2.1 Inputs
Detailed engineering drawings of the bottles – these were provided by the participating companies and remain
confidential. Glass Distribution – this was determined by actual measurements of the glass thickness. The details
and dimensions of the wine bottles are given in Table 1 and Appendix A: Figure shows the glass distribution of
the samples.
Table 1 Physical dimensions of wine bottles investigated.
Bottle ID
Dimensions
Average Glass Thickness
A
Weight
(g)
360
Height
(mm)
283
Diameter
(mm)
75
Body
(mm)
2.5
Shoulder
(mm)
1.5
Neck
(mm)
2.6
Base
(mm)
5.4
B
430
279
77
2.8
3.0
3.5
6.5
C
475
299
76
3.1
2.7
4.0
6.3
4.2.2 Outputs
Finite element analysis can be used to predict the performance of a container to a range of operating conditions.
For the purposes of this demonstration the bottles were modelled for:
Combined internal pressure and head loading
Containers are filled, capped or corked and then transported usually in
palletised form to the end user. Capping and palletising subject containers to
large vertical loads. For the purposes of the modelling exercise a loading force
was applied to the top of the container to represent the head load of 4020N
(410 kg) which is well in excess of the 1080 to 1370N (110 to 140 kg) typically
experienced by wine bottles finished with a cork or screw closure.
The model also simulated an internal pressure of 1.21 N/mm2. Pressure testing
is not a general requirement in the case of still wine but has been included
here to demonstrate the range of simulations that can be achieved with FEA.
Pressure testing is however needed for products such as sparkling wine, where
carbon dioxide is held in bottle under pressure.
Vertical Load due to Stacking
Pendulum impact at the shoulder
Containers are subject to impacts during filling, transportation and use. The
standard test comprises a simple pendulum which is used to deliver a
calibrated impact to the shoulder or heel of a container, these being the
natural contact points. For the purposes of the modelling exercise an impact of
100 cms per second was simulated in the shoulder location.
Industry Standard
impact tester
10
4.3
Results
Detailed results for all the modelling work are given in Appendix A. A single example of each simulation is
included in this section for the purposes of illustration.
The outputs are provided in the form of graphical presentation in which the predicted and target stresses are
plotted alongside the container’s profile.
4.3.1 Internal Pressure and Head load
This simulation replicates conditions of a uniform internal pressure of 1.209 N/mm2 (175 psi) and a simultaneous
vertical load of 4000N.
Details of the individual simulations are give in Appendix A and that for container A is also reproduced below
(Figure ) for the purposes of clarification. The output from the simulation shows the tensile stresses generated
on the outside and inside surfaces of the wine bottle. The blue line represents the tensile stress generated by
the simulated test conditions, and the red line represents the maximum allowable stress at that the bottle should
be designed to withstand (“target stress”). The target maximum stress set used assumes good handling, i.e. little
surface damage. Different areas of the container are given different allowable stresses, depending upon their
location and likelihood for damage. These areas, or “zones”, (e.g. the punt (bottle bottom), the shoulder and the
neck) are identified automatically by the software, and the appropriate allowable stress from the chosen set of
limits is displayed along that area.
Figure 7 Combined pressure and head load for wine bottle A.
Glass profile
Predicted
Stress
Failure
Indicator
Line
11
The position and the maximum stresses experienced by each of the three sample bottles under the simulated test
conditions are given in Table 2.
Table 2 Maximum stresses from combined pressure and head loading.
Bottle
ID
A
(360g)
B
(430g)
C
(470g)
Inside Surface
Outside Surface
Stress (N/mm2)
Position
Model
Safe
Limit
Heel
38
Heel
Heel
Stress (N/mm2)
Position
190
Safety
Margin
(%)
80
Model
Safe
Limit
Shoulder
35
135
Safety
Margin
(%)
74
47
190
75
Shoulder
24
135
82
34
190
82
Shoulder
18
135
87
As the bottles are all of a similar size, shape and glass thickness, a similar performance was to be expected under
pressure and head loading.
In general the results do show that thicker bottles are stronger, but the more important finding is that the lightest
bottle is still comfortably below the maximum allowable stress level and little extra strength is provided by the
addition of more glass. Indeed bottle B is seen to be more prone to failure at the heel than bottle A despite being
some 70g heavier; the additional glass not being used to strengthen the weak point.
Bottle A (Appendix A: Figure) was the lightest and has somewhat thin shoulders which are causing a stress peak
at this point on both inside and outside surfaces However, the stress is still below the acceptable level.
Bottle B (Appendix A: Figure ) has stress levels well below the acceptable limits in all areas but does show a
slightly high value at the heel which is consistent with the results from the physical pendulum testing (Appendix
B).
Bottle C (Appendix A: Figure 1) similarly shows stress levels well below the acceptable limits in all areas. This
bottle has a more pronounced punt, but the glass thickness around the heel is well distributed and the stress
levels do not show a large peak.
4.3.2 Impact Loading
The impact loading simulation was designed to replicate the action of the industry standard American Glass
Research (AGR) pendulum tester impacting on the shoulder region with a hammer speed of 100cms/s. The
simulated test set up is shown in Figure (the V-Bar is the part of the test rig against which the bottle sits as it is
being struck) and an example of the simulation output (bottle A) is shown in Figure .
12
Figure 8 Impact simulation conditions.
V-Bar Support
Impactor
@100 cms/s
Figure 9 Impact loading for wine bottle A.
Inside Surface: Section Angle = 0º (through impact point)
13
Outside Surface: Section Angle = 43º (through hinge point)
Details of the individual simulations are given in Appendix A Figure , Figure & Figure and that for container A is
also reproduced below (Figure ) for the purposes of clarification. The output from the simulation shows the
tensile stresses generated on the outside and inside surfaces of the wine bottle.
The output from the analysis are in the form of “stress maps” which show the maximum tensile stress generated
on the outside and inside surfaces of the containers during the impact incident (i.e. the stresses shown are the
maximum at any given point during the very short period of the strike, so would not in practice occur
simultaneously). The maximum tensile stress will be developed on the inside of the bottle immediately behind the
impact point. As the glass surface on the inside of the bottle should be in near pristine condition it should be well
able to absorb the impact energy. Pristine glass typically is able to withstand 650 N/mm2.
The stress map produced for the outside surface is of interest not only at the point of impact but also for the
position of any hinge points which are also possible failure points and which could be strengthened by small
design changes if necessary. It is prudent to assume that the outside surface will have suffered some wear and
tear and a value of around 70 N/mm2 is accepted as normal for lightly scratched surfaces.
The maximum tensile stresses predicted by the finite element analysis are given below in given in Table 3. The
results have also been compared with those of the physical testing in Figure and are seen to be consistent with
the thinner bottle developing higher stress levels for a given impact and thus faring less well with the pendulum
testing.
The full data sets are given in Appendix A and a summary of maximum stresses developed is given below in
Table 3.
14
Table 3 Maximum tensile stresses for impact stresses for the three wine bottles investigated.
Inside Surface (N/mm2)
Container
Outside Surface (N/mm2)
A (360g)
447
51
B (430g)
317
27
C (470g)
280
24
Figure 10 Comparison of physical testing and finite element analysis (shoulder strikes only).
Pendulum Results vs. Predicted FEA Maximum Stress
shoulder position only
500
Max Tensile Stress ((N/mm2)
450
400
350
300
250
200
150
100
100
150
200
250
300
350
Impact Resistance (cm/s)
The pattern of stress distribution is seen to be broadly the same for all the bottles. At the point of impact the
bottle suffers a concentrated region of stress on the inside of the bottle due to the impact, as that surface is the
one being stretched. On the outside of the bottles the highest stresses occur at the “hinge points” (see Figure 4)
which are nominally located on either side of the impact site. The actual hinge point positioning depends upon
the geometry, thickness and impact location.
The results of the analysis are seen to be in broad agreement with the physical testing with all maximum stresses
being within the acceptable range for a working bottle for both internal and external surfaces.
Bottle A (Appendix A: Figure was the lightest and displayed the highest tensile stresses on inner surface. The
impact was simulated at the shoulder location, where the glass was relatively thin and thus it was to be expected
that this container would not perform as well as the other two containers, which had thicker walls at this point.
Laboratory pendulum testing confirms this result (Appendix B). This container also showed the highest stresses
on the outside surface, again due to the fact that the hinge points fell within the thinner area of glass.
Two sets of hinge points are apparent in Bottle A. This is can be associated with containers having significant
curvatures at the shoulder. It can also be seen that principal hinge points occur at the top of the shoulder on
Bottle A.
Bottle B (Appendix A: Figure 16 ) has its principal hinge point somewhat lower in the bottle and it has a barely
discernable secondary hinge point coinciding with the upper shoulder region. This bottle performed least well for
stress in the heel region.
15
Bottle C (Appendix A: Figure ) has the largest average wall thickness and displays the least stress. The two sets
of hinge points are clearly defined, coinciding with the top and bottom of the shoulder region.
4.3.3 Physical Testing
Physical testing of the three bottles was undertaken by GTS. Testing was confined to impact testing and glass
distribution. As the bottles were production items they had already undergone extensive testing and were thus
known to be fit for purpose. The full data sets of the physical testing are given in Appendix B and a summary is
given below in Table 4.
Table 4 Summary of the physical impact testing.
Bottle ID
Heel impacts (cm/s)
Shoulder impacts (cm/s)
Minimum
Maximum
Average
Minimum
Maximum
Average
A (360g)
120
260
188
120
220
169
B (430g)
120
200
141
180
>320
263
C (470g)
200
240
223
170
>320
>290
The American Glass Research (AGR) pendulum tester is the industry standard and used worldwide but no formal
minimum standards are set for individual products, e.g. wine bottles, against which they could be judged on a
“pass/fail” basis. In practice glass manufacturers and testing laboratories are aware of the strength required by
individual products and are guided by experience in deciding a safe level. In the case of bottles to be used for
non-carbonated wine a minimum value of around 80 cm/s is considered to be an appropriate minimum.
Based on a minimum strength value of 80cm/s all three bottles are seen to be comfortably within the limit. The
results do display a broad correlation with weight with the heavier, thicker bottle C displaying the greatest
strength but nonetheless bottle A is perfectly serviceable and some 24% lighter than bottle C. The physical
testing also confirms some of the results of the FEA work by confirming that the design of bottle B has produced
a (relatively) weak heel and that the shoulder on bottle A would be somewhat weaker than the other bottles.
4.4
Practical Experience of Lightweighting Wine Bottles
During the course of the project a number of lightweighting initiatives were completed and bottles are now
available for use in the on and off trade. Ideally developing a new bottle will begin by the formation of a project
team which will include representatives of all the members of the supply chain including: the glass manufacturer,
the packer filler and the customer. The lightweighting process will then be able to progress rapidly as there will
be early and open involvement of all interested parties who will be able to ensure that the product meets all of
the customer’s requirements whilst not encountering any unforeseen difficulties during manufacture and filling
processes.
This approach adopted by the consortium that produced the lightweight bottle initially for use by Tesco and
latterly for the Co-op and proved to be very effective and contributed over 3,500 tonnes to the project’s target.
5.0
Other Comparative Studies of Lightweighted Container Strengths
A literature search revealed that other researchers have attempted to compare the performance of established
containers with their lightweighted equivalents. Jaime S.B.M. et al3 compared the mechanical performance of
lightweight glass packages produced by the Narrow Neck Press and Blow and the same regular weight glass
packages produced by the conventional process Blow and Blow.
3
Packaging Technology and Science, Volume 15, Number 4, July/August 2002, pp. 225-230
16
Several parameters including glass thickness distribution, mechanical performance (impact, vertical load and
thermal shock strength) and the performance of lightweight glass packages under transport simulation were
monitored. The study found that the lightweight glass containers had a more uniform thickness distribution in
comparison with the regular weight containers and a better performance (about 33% improvement) in relation to
the impact strength, especially in the heel, even when evaluated after line simulation.
The lightweight glass container also gave a better performance with respect to vertical load strength, both before
and after line simulation. Both containers withstood the temperature difference of 42°C that such packages are
supposed to resist according to thermal shock specification. However, due to the better thickness distribution of
lightweight glass packages, they were able to withstand a 5 to 10°C higher maximum temperature difference of
than the regular weight containers.
6.0
Conclusions
The weight of a wine bottle is not necessarily a good indicator as to its strength. A strong bottle will have the
glass well distributed; have no thin spots and few surface defects.
Glass is an inherently strong material but being brittle it is susceptible to failure when subjected to high tensile
stresses. Small surface defects act to weaken glass and are usually the focal point of breakages. Good
manufacturing methods that minimise these defects will produce a stronger bottle.
Most wine bottles are manufactured to designs appropriate to older forming technology and practices which could
now be revised to be lighter and yet sufficiently strong. Most glass manufacturers have added some lighter wine
bottles to their portfolio and these may be available if requested.
Some wine brand owners are still apprehensive about lightweighting, feeling that a move to a lighter version
would cheapen their image. However, others including Adnams and Coca Cola have used lightweighting to put
out a positive environmental message. Whilst lighter wine bottles are generally only marginally cheaper to
purchase than heavier versions they do reduce transport costs, and even small dimensional changes may
translate into an increase in the number each pallet can hold.
Modern manufacturing methods permit container manufacturers to produce wine bottles that are significantly
lighter than was previously possible, without compromising safety. Computational techniques such as Finite
Element Analysis (FEA) can reliably be used to ensure that new design concepts will produce a safe, strong wine
bottle without the need for extensive proto-typing. The technique is not able to predict the actual strength of a
container design but rather to determine which areas of a container are potential weak spots and some physical
testing will still be required.
The results of the FEA simulations of three nominally similar wine bottles were able to able to establish that the
designs would produce bottles with the required strength characteristics. The simulations were also able to
identify any (relative) weak points in the design which were confirmed by physical testing.
Remote sensing technology now offers the chance to monitor the performance of a container throughout the
supply chain and identify those areas in the system where containers are prone to damage.
New products can more rapidly be brought through the design, development and production stages if a guided by
project team with members drawn from all stages in the supply chain.
17
Appendix A: Modelling Results
The results of the finite element analysis conducted on the three bottles are given below.
Bottle Geometry
Based on technical drawings and the detailed wall thickness measurements made by GTS, Rockfield were able to
construct a glass profile for each item. Figure shows the profiles of these bottles.
Figure 11 Glass Distribution of the three wine bottles investigated.
Bottle A
Height (mm)
Diameter (mm)
Weight (g)
283
75
360
Bottle B
Bottle C
279
77
430
299
76
475
18
Internal Pressure and Head Loading
The results for simulated internal pressure and head loading are shown in Figures 12 to 14. The simulation
replicates conditions of a uniform internal pressure of 1.209 N/mm2 (175 psi) and a simultaneous vertical load of
4020N (410 kg). These graphs show the tensile stresses generated on the outside and inside surfaces of the
containers. The blue line represents the tensile stress generated by the simulated test conditions, and the red
line represents the maximum allowable stress at that the bottle should be designed to withstand (“target stress”).
The target stress set used assumes good handling, i.e. little surface damage. Different areas of the container are
given different allowable stresses, depending upon their location and likelihood for damage. These areas, or
“zones”, (e.g. the punt (bottle bottom), the shoulder and the neck) are identified automatically by the software,
and the appropriate allowable stress from the chosen set of limits is displayed along that area
Figure 12 Combined pressure and head load for container A (lightest weight).
19
Figure 13 Combined pressure and head load for container B (intermediate weight).
Figure 14 Combined pressure and head load for container C (heaviest weight).
20
Impact loading
Figure 15 Impact loading for container A
High
Stress
Stress
Point
Impact
21
Figure 16 Impact loading for container B
Stress
Point
Stress
Point
Impact
Outside Surface: Section Angle = 116.5º (through hinge point)
22
Figure 17 Impact loading for container C.
Stress
Point
Stress
Point
Impact
23
Appendix B: Physical Testing Data &
Results
Results of the physical testing carried out at the GTS technical facility are given below. Caution must be
exercised in making comparisons between modelled and physical data as the latter is greatly influenced by the
degree to which the test samples have been handled prior to the testing.
HEEL IMPACTS
SHOULDER IMPACTS
Bottle A
Bottle A
Sample
Number
Impact
Strength
(cm/s)
A1
A2
A3
A4
A5
A6
A7
A8
A9
A10
Min
Max
Average
Std Dev
150
150
240
170
230
150
170
120
260
240
120
260
188
49.4
Position of Failure
Mould Seam
90° To Mould Seam
Mould Seam
Mould Seam
Mould Seam
90° To Mould Seam
90° To Mould Seam
Mould Seam
Mould Seam
90° To Mould Seam
Sample
Number
Impact
Strength
(cm/s)
A11
A12
A13
A14
A15
A16
A17
A18
A19
A20
Min
Max
Average
Std Dev
150
140
170
120
190
220
180
160
180
180
120
220
169
28.1
Bottle B
Sample
Number
Mould
Number
B1
B2
B3
B4
B5
B6
B7
B8
B9
B10
Min
Max
Average
Std Dev
160
120
120
140
150
130
140
200
130
120
120
200
141
24.7
Position of Failure
90° To Mould Seam
90° To Mould Seam
90° To Mould Seam
90° To Mould Seam
90° To Mould Seam
Mould Seam
Mould Seam
Mould Seam
Mould Seam
90° To Mould Seam
Position of Failure
90° To Mould Seam
Mould Seam
90° To Mould Seam
Mould Seam
90° To Mould Seam
Mould Seam
Mould Seam
90° To Mould Seam
Mould Seam
Mould Seam
Bottle B
Sample
Number
Mould
Number
B11
B12
B13
B14
B15
B16
B17
B18
B19
B20
Min
Max
Average
Std Dev
180
280
250
>320
220
300
220
>320
280
260
180
>320
>263
>46.2
Position of Failure
90° To Mould Seam
90° To Mould Seam
90° To Mould Seam
No Failure
Mould Seam
Mould Seam
Mould Seam
No Failure
Mould Seam
90° To Mould Seam
24
HEEL IMPACTS
SHOULDER IMPACTS
Bottle C
Bottle C
Sample
Number
C1
C2
C3
C4
C5
C6
C7
C8
C9
C10
Min
Max
Average
Std Dev
Impact
Strength
(cm/s)
200
220
230
220
230
220
220
230
220
240
200
240
223
10.6
Mould Seam
90° To Mould Seam
90° To Mould Seam
90° To Mould Seam
Mould Seam
Mould Seam
Mould Seam
Mould Seam
90° To Mould Seam
Mould Seam
Bottle A
Min
Max
Average
Std Dev
Sample
Number
Position of Failure
C11
C12
C13
C14
C15
C16
C17
C18
C19
C20
Min
Max
Average
Std Dev
Bottle C
Impact
Strength
(cm/s)
170
300
310
>320
>320
>320
230
290
>320
>320
170
>320
>290
>50.6
Position of Failure
90° To Mould Seam
Mould Seam
Mould Seam
No Failure
No Failure
No Failure
Mould Seam
Mould Seam
No Failure
No Failure
Bottle C
HEEL
SHOULDER
HEEL
SHOULDER
HEEL
SHOULDER
120
260
188
49.4
120
220
169
28.1
120
200
141
24.7
180
>320
>263
>46.2
200
240
223
10.6
170
>320
>290
>50.6
.
It should be noted that as the bottles are of different designs and the natural contact (impact) points vary it is
difficult to draw direct comparison between the 3 bottle types.
25
Wall thickness measurements and check weight calculation
Bottle ID
A (Light)
Notes
Bottle weight
360 g
Glass thickness measurements made at 15mm intervals starting +15 mm from base
Also on 2 axes for base - 5 measurements each axis (includes centre = 9 measurement - centre common)
Q2
360
A (Light)
(Flint)
Body
Measuement
Point
A
B
C
D
base
15
30
45
60
75
90
105
120
135
150
165
180
195
210
225
240
255
270
crown
3.60
4.05
3.59
2.77
2.42
2.30
2.44
2.44
2.31
2.02
1.88
2.08
1.53
1.42
2.18
2.43
3.45
xxx
3.36
3.89
3.38
2.56
2.30
2.15
2.16
2.12
2.07
1.81
1.74
1.89
1.46
1.51
2.15
2.24
3.18
xxx
3.49
3.84
3.42
2.67
2.35
2.22
2.17
2.21
2.18
1.98
1.84
1.85
1.52
1.42
2.11
2.27
3.46
xxx
3.31
3.72
3.23
2.42
2.16
2.02
2.00
2.08
2.07
1.88
1.75
1.75
1.43
1.43
2.03
2.35
3.49
xxx
Bottle base measurements (mm)
5.26
5.79
centre
5.74
4.94
B >>D
5.46
5.6
5.66
A >> C
Bottle ID
B (Medium)
Notes
B >>D
Diameter
74.6
176.0
219.5
43.5
1.47
52.5
Neck
219.5
283.2
63.7
2.61
26
5.41
74.6
Body
Height
(mm)
176.0
Thickness
(mm)
2.5
Diameter
(mm)
74.6
Volume
(mm3)
99629
Wt
(g)
249
Shoulder
43.5
1.5
52.5
10219
26
Neck
63.7
2.6
26.0
12225
31
5.4
74.6
23649
59
Glass Volumes
Bottom
total
283
364
5.15
5.2
Bottle weight
432 g
Body
A
B
C
D
2.81
2.63
2.62
1.62
2.20
2.54
2.45
2.40
2.25
2.05
1.81
1.99
2.35
2.96
3.11
3.35
xxx
xxx
3.56
3.69
3.02
2.50
3.26
3.58
3.24
3.03
2.77
2.68
2.70
2.94
3.23
3.07
3.45
3.76
xxx
xxx
3.73
3.43
2.88
2.33
4.03
3.40
3.29
3.07
3.12
3.21
3.66
4.04
4.45
3.43
3.69
3.85
xxx
xxx
3.16
2.80
2.60
2.01
3.53
3.37
2.93
2.63
2.36
2.20
1.83
2.06
2.54
3.05
3.32
3.49
xxx
xxx
base
A >> C
Thickness
2.50
Bottom
Measuement
15
30
45
60
75
90
105
120
135
150
165
180
195
210
225
240
255
270
crown
Height
176.0
Shoulder
Wall thickness measurements (mm)
Point
Simplified calculation of weight
From
To
0.0
176.0
5.61
7.00
centre
7.05
7.06
6.02
7.05
6.93
From
To
0.0
171.0
Height
171.0
Thickness
2.84
Diameter
77
Shoulder
171.0
211.0
40.0
3.01
55.5
Neck
211.0
280.0
69.0
3.50
27
6.54
77
Bottom
Body
Height
(mm)
171.0
Thickness
(mm)
2.8
Diameter
(mm)
77.0
Volume
(mm3)
113177
Wt
(g)
283
Shoulder
40.0
3.0
55.5
19852
50
Neck
69.0
3.5
27.0
17842
45
6.5
77.0
30448
76
Glass Volumes
Bottom
Total
280.0
453
6.18
5.94
26
Bottle ID
Notes
C (Heavy)
Bottle weight
475 g
Rockware
475
3219
(Green)
Body
15
30
45
60
75
90
105
120
135
150
165
180
195
210
225
240
255
270
A
B
C
D
3.44
3.05
2.41
1.97
2.49
2.66
2.71
2.73
2.62
2.45
2.30
2.33
2.94
3.40
4.40
4.30
4.20
3.90
3.90
3.36
2.67
2.67
3.65
3.64
3.44
3.31
3.13
2.90
2.68
2.68
3.12
3.33
4.42
4.25
4.09
3.67
3.89
3.50
3.12
3.05
4.11
4.08
3.78
3.70
3.54
3.29
2.98
2.84
3.13
3.24
4.33
4.32
4.26
3.63
3.20
2.85
2.38
2.18
2.67
3.06
3.14
3.10
2.97
2.63
2.39
2.33
2.88
3.25
4.24
4.32
4.24
3.72
5.89
6.84
centre
6.56
5.92
B >>D
6.59
6.84
6.71
A >> C
From
To
0.0
152.4
Height
152.4
Thickness
3.09
Diameter
76.1
Shoulder
152.4
203.4
51.0
2.72
54.1
Neck
203.4
299.3
95.9
3.98
26
6.26
74.6
Bottom
Body
Height
(mm)
152.4
Thickness
(mm)
3.1
Diameter
(mm)
76.1
Volume
(mm3)
107893
Wt
(g)
270
Shoulder
51.0
2.7
54.1
22368
56
Neck
95.9
4.0
26.0
26383
66
6.3
74.6
27346
68
Glass Volumes
Bottom
total
299
460
5.85
5.10
27
www.wrap.org.uk/retail

Lightweight Wine Bottles

Transcription

Similar documents

Octatube Adverentie Challenging Glass 3

Lagaria Pinot Grigio KeyKeg Sell Sheet with Text Box

How to Read a Glass Bottle

LARGE ESTATE AUCTION

guaranteed guaranteed

DocuCentre SC2020 - Quick Reference Guide

Glass instructions

Angle duplication guide

Emailer _ glasspro India 2017.cdr

What is Message in a Bottle? How does it work?