fatigue tests

8/12/2019 Fatigue Tests

1/13

Chapter 12FATIGUE

Fatigue could be defined as any change in the properties of a materialcaused by prolonged action of stress or strain, but this general definitionwould then include creep and stress relaxation. Here, fatigue will be taken tocover only changes resulting from repeated cyclic deformation which means,in effect, long term dynamic testing.Subjecting a rubber to repeated deformation cycles results in a change instiffness and a loss of mechanical strength. In some products, even arelatively small change in stiffness can be important, but this measure offatigue is relatively little used, certainly not in standard tests. It would berelatively straightforward, although perhaps expensive in machine time, tocontinue a dynamic test as discussed in Chapter 9 over a very long periodand monitor the change in modulus. Alternatively, modulus could bemeasured at intervals after djmamic cycling on a separate apparatus. In manyproducts, notably tyres, it is the loss in strength shown by cracking and/orcomplete rupture which is considered to be the important aspect of fatigueand this is the measure of fatigue which is normally used in laboratory testson rubbers.The manner of breakdown will vary according to the geometry of thecomponent, the type of stressing and the environmental conditions. Themechanisms which may contribute to the breakdown include thermaldegradation, oxidation and attack by ozone, as well as the basic propagationof cracks by tearing. In rubber testing, it is normal to distinguish betweentwo types of fatigue test; tests in which the aim is to induce and/or propagatecracks without subjecting the test piece to large increases in temperature, andtests in which the prime aim is to cause heating of the specimen by thestressing process. The former type is generally referred to as flex-cracking orcut-growth tests and the latter as heat build-up. This division leaves out


2/13

246 Physical testing of rubberspecialised tests for particular products which may have characteristics ofboth types. For example, tyres fail by fatigue in which heat build-up isimportant and also suffer from groove and sidewall cracking, so that,logically, a realistic test would simulate both.

1. FLEX-CRACKING NDCUT GROWTH TESTSThe vast majority of flex-cracking tests strain the test piece in flexure,representing the mode of deformation experienced in service by suchimportant products as tyres, belting and footwear. Unfortunately, despite the

obvious logic of this approach, there are disadvantages. The principalproblem is that it is difficult to control the degree of bending , which may, forexample, vary with the m odulus of the rubber and, because the fatigue life ofrubber will be sensitive to the magnitude of the applied strain, misleadingresults may be obtained. The more nearly the deformation produced in thelaboratory test reproduces that experienced in service the better should bethe hope of correlation. It is hardly necessary to add that most products aresubjected to a most complex pattern of straining. The alternative approach isto use a simple but reproducible mode of deformation such as pure tension.1.1 Flexing methods

A variety of flex tests have been used, many intended for particularproducts such as belting, footwear and coated fabrics, but a number of them,although once well known, are not now in such common use. A usefulreview was given by Buist and Williams in 95 ^Four types of machine in which bending is produced in different waysare shown schematically in Figure 12.1. The most widely known andstandardised apparatus, the De Mattia, has the action shown in Figure12.1(a). The test piece, which is a strip with a transverse groove, is fixed intwo clamps which m ove towards each other to bend the strip into a loop. Themaximum surface strain at the critical point X is somewhat indeterminate. Inthe 'flipper' or Torrens machine (Figure 12.1(b)), the strip test piece, fixed ina slot in the periphery of a rotating wheel, is bent against a fixed, but freelyrotatable, roller. Again, the radius of curvature, and hence the maximumsurface strain, is not precisely controlled. In the Du Pont machine (Figure12.1(c)), the test pieces are connected together to form an endless belt andrun over a series of pulleys of specified diameters. Although the overallradius of bending is controlled, the surface strain in the test piece iscomplicated by it having several transverse 'V grooves. The Ross machine(Figure 12.1(d)) bends the test piece through 90"^ over a rod and the


3/13

Fatigue 247maximum surface strain is rather more controlled than in the other machinesdescribed.

The control of minimum strain is even more important than the control ofmaximum strain in a flexing cycle because rate of cracking is muchincreased if this is zero. In all the methods described above, the strain isdeliberately intended to be zero, but probably only in the Ross apparatus isthis achieved precisely and in a reproducible manner. In all bendingmethods, the maximum strain depends on the thickness of the test piece and,hence, this must be closely controlled.

ud) L^-Figure12-1.Types of flexing test, (a) De Mattia; (b) "flipper" machine; (c) Du Pont;

(d) Ross. In (a), (b) and (d) the flexe d form is shown by broken Hnes.There is now one international standard for flex testing, ISO 132^ whichcovers both tests for crack initiation and for cut growth. Previously, thesewere separated into two standards. The different significance of the two


4/13

248 Physical testing of rubbertypes of test is illustrated by the fact that natural rubber fairly quicklydevelops fine cracks in a flex-cracking test but is relatively resistant to thefurther growth of these cracks or of a purposely made cut, whereas SBRshows just the opposite behaviour. For both methods, the standard specifiesthe De Mattia apparatus and the same test conditions, the essential differencebeing that for cut growth tests a cut is made through the bottom of thegroove in the test piece before flexing is started.The De M attia apparatus operates at 5 Hz with a maximum separation of75 mm between the grips and a travel of 57 mm (i.e. between 75 and 18mm). The test piece is a strip with a moulded groove which is intended toconcentrate stress, and hence cracking, in the centre of the strip. The testpieces are inserted into the grips at maximum separation to give zero strain.In flex-cracking tests, one ofthemost difficult problems is how to assessthe degree of cracking. Visual examination is the only really feasibleprocedure and, inevitably, the assessment is subjective and operatordependent. Unfortunately, the pattern of cracking in a De Mattia test varieswith the type of rubber and is likely to start at the edges of the test piece,although this can be virtually eliminated by radiusing the edges. Alternativegrading systems were discussed by Boss and Greensmith^ and the 'modifiedcode' they suggested is now essentially the procedure specified in ISO 132.It is based primarily on the length of the largest crack present at any stageand the depth of the crack is ignored. Any more complicated processinvolving the measurement of length, depth and number of cracks isgenerally unacceptable and, in any case, any precision gained is usuallymasked by between test piece variability. Judging against a standard set ofphotographs is only of any use if the rubber under test follows the samepattern as that illustrated. Judging on the time to the first appearance ofcracks gives only a single point measurement and is liable to be morevariable than taking a series of grades of increasing severity.

In the cut-growth method, a 2 mm cut is made through the rubber, andthe geometry ofthechisel-like piercing tool is given in detail. The length ofthe cut is measured at intervals and the number of cycles for it to increase by2 mm, 6 mm and 10 mm is deduced.The British Standard is identical and numbered as BS ISO 132. ASTMhas several flexing fatigue methods and it can be seen from the lack ofuniformity in the titles that they were not developed as a group.ASTM D430'^ specifies three machines for flex-cracking. The ScottFlexer is included specifically to test for ply separation of composites suchas belting or tyres and is not used for rubber alone. It is a somewhat bulkyapparatus in which the test pieces are reciprocated over a rotatable hubwhilst held in tension. The arc of contact is about 165^ but was at one timewas erroneously given as 135^, resulting in at least one out of spec machine


5/13

Fatigue 249in the UK. Different diameter hubs are used for belts and tyres, 31.7 and14.3 mm respectively.The De Mattia method is not exactly the same as the ISO and BSmethods. The distance between the grips reciprocates between 19 and 75.9mm, whilst the test piece length between the grips is 76.2 mm. Hence, zerostrain is not quite reached. In addition to using the machine to repeatedlybend the test piece, a procedure is given in which a dumb-bell is cycled inpure tension. In both cases, the degree of cracking is judged by visualobservation and graded on a scale of 0 - 10, where 0 represents no cracksand 10 is essentially complete failure. The use ofthedirect tension mode ofstraining will be considered in Section 1.2.The third machine specified is the Du Pont Flexer, as briefly describedabove. The belt is made up of 21 test piece links which run over anarrangement of four pulleys under a tension of about 76N..ASTM D813^ specifies the De Mattia apparatus for cut-growthmeasurements and suggests that it should be used for materials which do notreadily initiate cracks when tested by the methods given in ASTM D430.The fact that some materials are difficult to initially crack but will readilypropagate tears is obviously of great practical importance, but it is verydebatable whether the two forms of test should be considered on an either/orbasis. The procedure and expression of results are not identical with ISO133.Also, rather oddly, the free length is 75.9 mm and, hence, differs fromASTM D430 for flex cracking.A second method for cut growth using the Ross Flexer is given in ASTMD1052^ but there appears to be no cross reference between this and D813.As discussed above, the Ross has the particular advantage of controlling themaximum and minimum strains rather more precisely than in other bendingtests. Because of this, it is a little surprising that the method is not morewidely used. In Britain, the Ross has been used to test soling materials forfootwear. To prevent heat build-up in the test piece, the apparatus operatedat a slower speed than in the ASTM standard and common practice was totest at 0C or -5^C.ASTM has also standardised a 'flipper' type of machine^ which usesgrooved and pierced test pieces to measure crack growth. The gap betweenthe revolving disk and the deflector bars can be varied so that the angle ofdeflection and, hence, the severity of test can be varied. The apparatus incontained in an oven so that tests can be made over a range of temperatures.1.2 Tests in Tension

All of the bending methods are to some extent arbitrary as to the degreeof strain used and in most tests neither maximum nor minimum strains are


6/13

250 Physical testing of rubberwell defined. By cycling in simple tension, strains can be reproduced moreeasily and a range of strains and prestrains can be readily realised with oneapparatus. A standard procedure for fatigue in tension adopted by ISO aroseout of the MRPRA work on the concept of tearing energy (see Chapter 8,Section 9).In this fracture mechanics approach, the rate of crack growth is a functionof the maximum value of tearing energy attained during the fatigue cycle.For a strip with an edge or central cut cycled in tension, a relation betweencycles to failure and the initial cut size for the case where that cut isrelatively very small can be derived^' :

N= n \{IKWYiC,where: N = fatigue life in cycles to failure, G = cut growth constant, K =function of the extension ratio, W = strain energy per unit volume, Co =initial depth of cut (or intrinsic flaw) and n = strain exponent dependentupon the nature of the polymer.Hence, at a constant value of K, a plot of log (N) against log (W) willhave a slope n. The value of n has been found to be about 1.5 for a naturalrubber tyre tread and 3 for an SBR tread. If no artificial cut is introducedthen Co is the effective size of a naturally occurring flaw. The strain energydensity, W, can be found from the area under the stress/strain curve for thetest piece and is strain dependent. The fatigue life is independent of thespecimen geometry when expressed in these terms. At low strains, theequation does not adequately describe the fatigue behaviour and there is afatigue limit corresponding to tearing energy, below which there is virtuallyno cut growth and fatigue life becomes very long (unless ozone crackingoccurs).Test can be made at a number of extensions and compounds can becompared in terms of fatigue life at the same strain or at the same strainenergy. In the latter case, absolute comparisons can be made on compoundsof different modulus. When comparing different rubbers, it is necessary totest at a number of strains or to define the severity of conditions which willoccur in service, because with the number of variables (G, K, W, n and Co)the ranking order may vary with the maximum strain employed.

ISO 6943^^ for fatigue in tension specifies two different types of testpieces, rings and dumb-bells, which correspond to the geometries used oncommercially available apparatus. There is, in principle, little differencebetween the two forms of test piece but dumb-bells are necessary forstudying directional effects. They are also easier than rings to cut from sheet,but normally a specially moulded sheet is required such that the dumb-bells


7/13

Fatigue 251have a raised bar across each tab end to aid location and gripping. There areno gripping problems with rings and the roller separation is a direct measureof strain.

The dumb-bells specified are the same as those for tensile stress/straintests except the preferred thickness is 1.5 mm. The ring is also the same asthe tensile ring, which means that the bulk ofthe two types of test piece aredifferent.The range of frequency specified is between and 5 Hz and the standardonly covers strain cycles passing through zero, although prestrains could beapplied. It is suggested that at least five test pieces should be tested at eachstrain and that usually it is desirable to test at a number of maximum strains.The strain on ring test pieces is calculated on the internal diameter (seeChapter 8, Section 5.1). The test is continued until complete failure of thetest piece occurs and then the number of cycles recorded.During the course of a test, the stress/strain relationship ofthe test piecewill change and there will also be a degree of set. It is recommended thatboth these parameters are measured at intervals and the results reported aswell as the fatigue life. The results can be presented in graphical form as log(fatigue life) against strain, log (strain energy density) or log (stress). Anannex gives explanatory notes, including a section on interpretation ofresults which introduces the concept ofafatigue limit.It is generally found that the relationship between 'fatigue life' andapplied stress or strain is of the form shown in Figure 12.2. The importantfeature of this so called S-N or Wohler curve is that, on reducing the stressor strain towards a particular value, the fatigue life increases virtually toinfinity, giving rise to the concept of a limiting fatigue lifeThe British standard^ is identical to the ISO and ASTM D4482^^ isessentially similar but only covers dum bbell test pieces.Fatigue life is influenced by the environmental conditions under whichthe test is carried out, in particular temperature, oxygen and ozone. Theeffects of these have been discussed by Derham et al*^ and Clapson andDove^"^, the latter authors also giving examples of the application of thetensile form of fatigue testing to practical applications.The publication of the ISO and equivalent standards was expected toencourage more workers to apply the fracture mechanics approach whichunderlies them to the prediction of fatigue in rubber products, although thestandard itself does not in fact go into the fracture mechanics theory. Judgingfrom the very considerable volume of literature that has been generated,there has been success in this direction.Examples of accounts of applying the fracture mechanics approach to thefatigue behaviour of rubbers are given in references 15-19. In a review oftesting methodology for reinforced rubber composites^^ it is concluded that


8/13

252 Physical testing of rubberthere can be coexistence of fracture mechanics and the probabihsticapproach of S-N curves. Fukahori^^ has demonstrated the generation ofanS-N master curve by superimposing data obtained at different crack lengthsand has analysed the application of the curves with fracture mechanics.

Fatiguelife

ULTIMATEFATIGUESTRENGTH

Stress or strain

Figure12-2.Relationship between fatigue life and applied stress or strain (S-N or Wohlercurve)The dependence of fatigue life on the maximum strain and, particularly,the minimum, non-zero strain has been demonstrated^^ using rather unusualcylindrical dumb-bells. A technique was developed^^ for estimating theeffect of stress relaxation on tearing energy measurements and for adjustingthe measured values for relaxation. This work also made tests on test piecesof different widths as being equivalent to increasing crack length.Eisele et al^"^ describe the so called tear analyzer using a strip test piecewith a cut in one edge cycled in tension, which can be considered the classic

geometry for obtaining fracture mechanics data on rubbers. Thissophisticated instrument introduces nothing new in concept but has atemperature controlled cham ber and can operate at different frequencies,pre-strains and strain amplitudes, with automatic compensation for set.


9/13

Fatigue 253Examples of it use are given by Kelbch et al^^ Sumner et al^^ and BoehmandStruve^^'^lYang and Kang^^ give a relation for estimating the crack length from thevariation in peak loads under constant strain displacement cycles. Sun et al ^investigated the applicability of Miner's rule to rubbers fatigued in tensionand determined the effect of loading sequence on crack growth.1 3 Other tests

The British standard guide to application of rubber testing to finiteelement analysis^ covers fatigue tests and suggests nine different possibletest piece geometries:- edge crack in a tensile strip, central crack in tensilestrip,pure shear test piece with either long edge crack or short central crack,angled test piece, trouser test piece, split test piece, simple shear, a twisteddisc,peel and rod pull out. The edge crack tensile geometry is most commonas the crack is easy to form and growth can be monitored optically (a centralcrack is more difficult to introduce). The particular characteristics of theother geometries are summ arized in the standard.Muhr and Thomas^^ suggest the angled test piece to eliminate the need toknow the strain-energy of the material, whilst Samsuri^^ et al developed aspecial equipment needed for the split tear test. Fleischman etal " developeda method using a disc in torsion with a circumferential pre-crack to simulatebelt edge interlaminar shear. Stevenson^^ considers the use of compressiontest pieces and the application of the results to offshore platform supportsand antivibration mountings. Takeuchi et al^^ describe a test speciallydeveloped for vibration insulators which uses a cylindrical dumb-bell shapedtest piece with ends bonded to metal. The deformation combines tension andcompression.An investigation of the influence of loading conditions on fatigue hfe^^involved a test piece with a rigid insert and a cut at the rubber/insertinterface. Liu et al^^ established a fatigue testing system to give periodicloading rather than continuous extension cycling and considered theadvantages of this approach. The connection between fatigue lives ofrubber/fabric materials in uniaxial and biaxial tension was established over atemperature range of up to 140^C at low frequencies^^. Saintier etal" ^studiedasymmetrically notched samples in a uniaxial machine and in torsion, andmade comparisons with finite element analysis.

Nechiporenko"^^ used a novel approach of repeated impact testing toobtain fatigue life under circumstances that would be relevant to tyres inquarries. He used a Schob pendulum with automatic delivery and countingof impacts, and needed only tests at two energies with up to 150 blows togive a rapid test.


10/13

254 Physical testing of rubberExplosive decompression in pressurized hose and seals can result indamage only after several decompression cycles, i.e. as a result of fatigue.Tests were made"^^ which produced fracture surfaces similar to those from

explosive decompression and the importance of maximum strain,temperature, void size and void position was highlighted.

2 HEAT BUILD-UPIt is rather confusing that the 'heat build-up' type of fatigue test is carriedout on an apparatus generally called a flexometer, which brings to mind

flex-cracking and cut-growth tests. The term heat build-up is not in fact aparticularly good one as rupture of the test piece, set and changes in stiffnesscan also be measured, but it serves to distinguish the tests from those whereonly surface cracking is of interest and the test piece geometry is such thattemperature rise is minimised.Flexometers or heat build-up fatigue apparatus operate in compression,shear or a combination of the two and various designs have been in use andstandardised, particularly by ASTM, for many years. The test piecegeometry and deformation cycle used are, inevitably, somewhat arbitraryand this perhaps contributed to it being much later before there was aninternational or British standard method.The international Standard, ISO 4666"^^, has the title 'Determination oftemperature rise and resistance to fatigue in flexometer testing', and is splitinto three parts: basic principles, the rotary flexometer and the compressionflexometer. The first part of ISO 4666 attempts to describe the basicprinciples of fatigue testing to give guidance on the interpretation of resultsusing particular apparatus and test conditions. This information would seemto be very necessary because results obtained under any particular conditionsare quite arbitrary and have no significance apart from the conditions used.Most fatigue tests apply a fixed pre-stress or strain, partly becausewithout bonding of the test piece it is necessary to hold the rubber in place.The amount of pre-stress or strain will affect the fatigue life; in particular thefatigue life is appreciably shortened ifthecyclic deformation passes near toor through zero strain.A fatigue test can be made with cycles of either constant strain am plitudeor constant force amplitude. With constant strain, the resultant stresses aregreater for higher stiffness rubbers, so that these are stressed more highlyand, other things being equal, will develop more heat. On the other hand,with constant stress a stiff rubber will deform less and consequently tend togive a better fatigue performance. Clearly, to avoid conflicting results it isnecessary to choose conditions which correspond with those met in service.


11/13

Fatigue 255It is also possible to use cycles of constant energy which is quite commonlythe situation met by such products as dampers and shock absorbers, but thisis more difficult from the apparatus point of view. The 'pre-stress' can in factalso be a constant stress or a constant strain.ISO 4666:Part remarks on the care necessary in measuring temperaturerise and the fact that the result depends on where the temperature ismeasured and on the test piece geometry"^" . It recomm ends testing at a seriesof strain or stress levels because a comparison of rubbers at one level onlycan be misleading. The standard also mentions the measurement of creepand set in the test piece after periods of dynamic cycling.The ASTM standard is 0623"^^ and has the somewhat confusing title ofHeat generation and flexing fatigue in compression. It specifies two types ofapparatus, the Goodrich flexometer and the Firestone flexometer. Thecompression flexometer of the ISO standard is essentially the same as theGoodrich and operates by superimposing a cyclic compression strain ontothe static deformation caused by a constant force. For the Goodrichflexometer test given in ASTMD623,the 17.8 mm diameter by 25 mm hightest piece is cycled at 1800 cycles/min with a stroke of4 45mm for 25 minand the temperature rise recorded. A choice of three static loads is given,alternative strokes suggested and two ambient temperatures, 50C and100C, recommended. Apart from measuring temperature rise, the staticdeflection of the test piece, its dynamic deflection, compression set, andindentation hardness are recorded. The ISO method has the same test piece,speed and choice of strokes, with the option of two pre-loads. Thetemperature rise, creep, compression set and fatigue life are reported.The ISO rotary flexometer and the Firestone both operate bysuperimposing a cyclic shear deformation onto a static compressivedeformation but the cyclic action of the two machines is not the same. TheISO apparatus is derived from the St Joe flexometer, which at one time wasincluded in the ASTM standard (up to 1962).The Firestone flexometer method in D623 is not very specific. Thestandard test pieces are in the shape of a frustum of a rectangular pyramidbut the use of any suitable shape is permitted when cut from products. Theapparatus operates at 800 cycles/min and a range of compression loads andamplitudes of oscillation are possible, but no particular conditions arespecified. The test piece is fatigued until a definite, but unspecified, decreasein the height of the test piece is reached, which is supposed to represent theonset of internal porosity. Parameters such as temperature rise and changesin compression are reported.The specification oftheISO rotary flexom eter is not m uch better. It usescylindrical test pieces and operates at 14.6 or 25 Hz. The axial compressioncan be either constant stress or constant strain and loading conditions are


12/13

256 Physical testing of rubbersuggested for both measurements of temperature rise and resistance tofatigue breakdown. Breakdown is not precisely defined. The vagueness ofthe ASTM and ISO methods for rotary flexometers reflects the arbitrarynature of these tests.BS903:Parts A49^^ and A50^^ are identical to ISO 4666:Parts 1 and 3.There is no British equivalent to the rotary flexometer, simply because suchan arbitrary apparatus was not considered worthy of standardisation and it isnot used in the UK.A draft is now being progressed in ISO TC 45 for a fourth part of ISO4666 to cover constant stress flexometers. There is no doubt that thereshould be a method to allow testing at constant stress but perhaps the mostencouraging thing about this new work is that the apparatus is based onmodern instrumentation and not on some arbitrary and historical mechanicaldevice. Unfortunately, at the time of writing, the draft needs a great deal ofwork before it could be called satisfactory.A servo hydraulic dynamic fatigue machine can, clearly, be used fortesting under constant strain as well as constant stress conditions and thiswould be preferable to the traditional mechanical instruments. Interestingly,a comparison has been made of dynamic mechanical properties measured ona new version of the Goodrich flexometer and a servo hydraulic tester"^^.

REFERENCES1. Buist J M and Williams G E. India Rubber World, 127,1951,p32 0,44 7 and 567.2.ISO 132, 1999. Determination of flex cracking and crack growth (De Mattia).3.Boss J D and G reensmith H W. J. IRI, May/June 1967,p.165.4.ASTM D430, 1995 (2000). Rubber deterioration - Dynamic fatigue.5.ASTM D813, 1995 (2000). Rubber deterioration - Crack growth.6. ASTM D1052, 1985 (1999). Measuring rubber deterioration - Cut growth using Ross

flexing apparatus.7. ASTM D 3629, 1999. Rubber property - Cut growth resistance.8. Gent A N, Lindley P B and Thomas A G. J. Appl. Polym. Sci., 8,4 55 ,19 64 .9. Lake G J and Lindley P B . Appl. Polym. S ci., 8, 707, 1964.10.ISO6943,1984. Determination of resistance to tension fatigue.11.BS 903 PartA51,1986. Determination of resistance to tension fatigue.12.ASTM D 4482, 1999. Rubber property - Extension cycling fatigue.13.Derham C J, Lake G J, Thom as A G . J. Rubb. Res. Inst. Malaya, 22, 2, 191, 1969.14.Clapson B E and Dove R A. Rubb. J. No. 12,41,Dec. 1972.15.Breidenbach R F, Lake G J. Phil. Trans. R. Soc. Lond. A 299,1981,pl89.16.Lake G J. Rubb. Chem . Technol., 68, 3,1995,p435.17.Young D G, Doyle M J. Antec 92 , 3-7 '' May 1992, Proceedings, Vol. 1,pl211.18.Bauman J T. ACS Rubber Division Meeting, Grand Rapids, 17-19* May 2004, Paper 5.19.Zhbakov B I. Int. Polym S ci. Technol., 23 , 5,1996,pT39.20. Causa A G, Borowczak M, Huang Y M. ACS Rubber Division Meeting, Cleveland, 21-

24'^October 1997, Paper 46.


13/13

Fatigue 25721 .Fukahori Y. Int. Polym. S ci. Techno ., 13,3, 1986, pT37.22.Alshuth T, Abraham F . Muanyag es Gummi, 40, 10,2003, p353.23 . Bauman J T, Goolsby R D. ACS Rubber Division Meeting, Anaheim, 6-9^^ May 1997,

Paper 70.24.Eisele U, Kelbch S, Engels H-W. Kaut. u Gummi Kunst., 45 , 12, 1992, pl 06 4.25 . Kelbch S A, Eisele U G, Magg H. ACS Rubber Division Meeting, Pittsburgh, 11-14^*^October 1994, Paper 19.26.Sumner A J M, Kelbch S A, Eisele U G. ACS Rubber Division Meeting, Cleveland, 17-20*October 1995, Paper 67.27.Bohm D , Struve J. Tire Technology Int., 2000, p55 .28.Bohm D , Struve J. Tire Technology Int., 2000, p67 .29.Yang K J, Kang K J. Rubb. Chem. Technol., 75 ,4 2002, p657.30.Sun C, G ent A, Marteny P. Tire Sci. and Technol., 28, 3, 2000, pl 96 .31 .BS 9 03- 5,20 04. Guide to the application of rubber testing to finite element analysis.32.Muhr A H, Thomas A G. ACS Rubb. Div. Cincinnati, 1988, Paper 32.33.Samsuri A, Piah M B M. J. Rubb. Research, 7, 2,20 04, p i15.34.Fleischman T S, Kerchman V, Ebbott T G . Tire Sci. Technol., 29, 2,2001,p91.35.Stevenson A. Polym. Test., 4, No. 2 - 4, 1984.36.Takeuchi K, Nakagawa M, Yamaguchi H, Okumotot T. Int. Polym. Sci. Technol., 20, 10,1993,pT64.37.CharrierP .Ostoja-Kuczynski E, Verron E, Gom et L, Chagnon G. IRC 2002, Prague, 1-4*July 2002 , Paper81.38.Liu Y, Wan Z, Tian Z, Du X , Jaing J, Yao M. Tire Sci. Technol., 27, 1, 1999, p48.39.Borodko T V . Int. Polym. Sci. Technol., 5, 3,1978,pTl .40 . Saintier N, Andre N, Cailletaud G, Piques R. IRC'98, Paris, 12-14* May 1998,proceedings pi 89 .41 .Nechiporenko A G. Int. Polym. Sci. Techn ol, 11 ,2, 1984, pT6.42. Edmond K, Ho E, Flitney R, Groves S, Embury P, Rivereau J M. 17* InternationalConference on Fluid Sealing, 8-10* , York, April2003,Proceedings, p2 41.43. ISO 4666 Parts 1-3, 1982. Determination of temperature rise and resistance to fatigue inflexometer testing.44 .Buckley D J. ACS Div. of Rubb. Chem. M eeting, May 1974, Toronto, Paper 34.45 .ASTM D 623,1999. Heat generation and flexing fatigue in compression.46. BS 903:Part A49, 1984. Determination of temperature rise and resistance to fatigue in

flexom eter testing (basic principles).47 . BS 903:Part A50, 1984. Determination of temperature rrise and resistance to fatigue inflexom eter testing (compression flexom eter).47 .Askea D W. ACS R ubber Division M eeting, Cleveland, 17-20* October 1995, Paper 100.

fatigue tests

Documents