Hi,
Are there any established SN curves or alike for natural rubber.
Are there any established SN curves or alike for natural rubber.
Tek-Tips is the largest IT community on the Internet today!
Members share and learn making Tek-Tips Forums the best source of peer-reviewed technical information on the Internet!
-
Congratulations JAE on being selected by the Eng-Tips community for having the most helpful posts in the forums last week. Way to Go!
SN curve for Natural Rubber 4
- Thread starter SOLIDMODEL1
- Start date
-
1
- #3
SN curves are fatigue-life curves (S=stress, N=cycles).
No, there are no SN curves for rubber. Standard fatigue theory deals with granular crystalline structures (metals). Although some people have tried to apply such theories to polymeric and elastomeric materials, the general consensus is that such an approach is not very fruitful.
There is a whole area of research trying to develop fatigue theories that are more appropriate to chained structures, but this is very much still in the theoretical/research stages.
Brad
No, there are no SN curves for rubber. Standard fatigue theory deals with granular crystalline structures (metals). Although some people have tried to apply such theories to polymeric and elastomeric materials, the general consensus is that such an approach is not very fruitful.
There is a whole area of research trying to develop fatigue theories that are more appropriate to chained structures, but this is very much still in the theoretical/research stages.
Brad
- Thread starter
- #5
SOLIDMODEL1
Mechanical
Cheers Brad, it's as I thought.
There's an interesting couple of curves in P.B.Lindley showing the fatigue life of NR subjected to tensile deformations at 2Hz. One at 0% minimum strain, maximum strain varying. The other at 250% maximum strain, minimum strain varying. The latter showing an increase in fatigue life as the minimum strain increases. I wonder if you could substitute the strains for stress from a stress/strain tensile test curve, so constructing an approximate SN curve.
There's an interesting couple of curves in P.B.Lindley showing the fatigue life of NR subjected to tensile deformations at 2Hz. One at 0% minimum strain, maximum strain varying. The other at 250% maximum strain, minimum strain varying. The latter showing an increase in fatigue life as the minimum strain increases. I wonder if you could substitute the strains for stress from a stress/strain tensile test curve, so constructing an approximate SN curve.
Even if SN methodology is valid, one must bear in mind that the fatigue performance of natural rubber is compound dependent. The fatigue life for any given stress/strain combination (or tearing energy if you like that approach) will be affected by the cure system used to cure the compound, the amount of carbon black and the type of carbon black in the compound, the amount of plasticizer, the type of antidegradant system, the presence of absence of minor amounts of other polymers such as polybutadiene or SBR..... So natural rubber as the polymer is only the starting point for the compound and fatigue results can be varied of several orders of magnitude based on the rest of the formulation.
Problem with with S-N curves is since cracks initiate from some flaws in the sample, large scatter is found in fatigue lives of specimens without a defined precrack. Therefore these tests are very tedious, time consuming and information obtained is specific to the specimen used which may not be representative. Having a pecrack in the sampe is more reproducible. This is called fatigue crack propagation. Tests can be load controlled or strain controlled. You should bear in mind an interesting feature of NR which is called strain induced cristalization. That causes its enhanced fatigue properties under non relaxing conditions (varied minimun strain), expaining Lindley results
Oscar
Oscar
rubberdoc--
The same holds true for any material in fatigue--fatigue properties are far more varied than "mechanical" properties such as moduli, density, etc.
But again, this is somewhat of a moot point as many have tried to apply conventional metal fatigue theories to rubbers and seen little success.
Brad
The same holds true for any material in fatigue--fatigue properties are far more varied than "mechanical" properties such as moduli, density, etc.
But again, this is somewhat of a moot point as many have tried to apply conventional metal fatigue theories to rubbers and seen little success.
Brad
-
1
- #9
Hi gents,
SN curves for rubber are usually referred to as Wohler (pronounced 'verler') curves. These are plotted with log cycles on x axis, against stress (n/mm2) on the Y axis.
These are experimentally derived for different polymers, by performing a mumber ( 6-10)of cyclic fatigue tests at different stess levels.
Most manufacturers of rubber antivibration mountings ( including the one I work for) routinely produce these curves, so the data should be available. The book 'Rubber Springs Design' by E.F. Gobel ( probably put of print) discusses this type of S/N curve.
Best regards
Tom
SN curves for rubber are usually referred to as Wohler (pronounced 'verler') curves. These are plotted with log cycles on x axis, against stress (n/mm2) on the Y axis.
These are experimentally derived for different polymers, by performing a mumber ( 6-10)of cyclic fatigue tests at different stess levels.
Most manufacturers of rubber antivibration mountings ( including the one I work for) routinely produce these curves, so the data should be available. The book 'Rubber Springs Design' by E.F. Gobel ( probably put of print) discusses this type of S/N curve.
Best regards
Tom
- Thread starter
- #10
SOLIDMODEL1
Mechanical
Thank you all for your input.
Tom,
Are you the same Tom Aspin that worked in Silentbloc Technical Dept.![[wavey2]](/data/assets/smilies/wavey2.gif "[wavey2] [wavey2]")
I was a mere test engineer back then, I'm now in the Technical Dept. along side Eddie Deeming. You would have known me as Colin (long story). Anyway, if you read this and feel inclined to do so, e-mail me at conrad.hextall@silvertown.co.uk
And yes, we do have the book you mentioned, silly of me to overlook it.
Regards,
Conrad.
Tom,
Are you the same Tom Aspin that worked in Silentbloc Technical Dept.
I was a mere test engineer back then, I'm now in the Technical Dept. along side Eddie Deeming. You would have known me as Colin (long story). Anyway, if you read this and feel inclined to do so, e-mail me at conrad.hextall@silvertown.co.ukAnd yes, we do have the book you mentioned, silly of me to overlook it.
Regards,
Conrad.
Ok, so I gather that S/N curves are not considered a good way to characterize how rubber products survive under cyclic loading. How do I go about describing how a rubber product will survive? I have a hydraulic o-ring that is subjected to various pressure loads throughout its life. In some applications, this o-ring is not adequately supported and under pressure can begin to extrude into another cavity. This is not a one-time overload that causes the o-ring to fail. It is more of a cyclic loading that eventually fatigues the material, the o-ring fails and I have a leak. I am attempting to characterize the behavior of the o-ring so I can predict how many of these o-rings will eventually fail and when. A S/N approach does not seem to be giving me reasonable results. I need to find another approach. Any help is appreciated.
-
2
- #12
. . . as was already mentioned earlier in this thread, an attempt to quantify the service life of a rubber component (in the present case a seal)one is confronted with several limitations concerning a proper methodology and lack of knowledge on the fatigue processes of rubbers.
Despite of these limitations I would suggest the following procedure to follow in order to receive a prognoses for service life of the o-ring in question:
Perform a non-linear finite element analysis of your sealing problem, where a typical cyclic loading should be simulated. Detect the areas of the cross section where strains are large. Get the full information on the strain tensor in those areas. Determine the components of the strain tensor which are in tension, because they may be the most dangerous ones. If you have the impression that one of the tension-strain components is much greater than the other ones (which often appears to be)you may be justified to handle that deformation mode for further considerations only. Now correlate the strain energy density found in the considered area of the cross section with the tear energy appropriate for the given deformation mode (see for example "How to design rubber components" by A. Gent). Thus you will get a relation dealing with a virtual crack you think about to be present in the considered area of the cross section of the o-ring. Then you may assume you would know the critical tearing energy ( ... and indeed one can assume a "most-lower" value - to be on the safe side). With all those assumption in mind you can state whether the current dynamical loading may be critical concerning a possible growth of some virtual cracks (flaws etc of certain crack-length). If you thus come to the conclusion that initial cracks ( always be present, independently of compound quality which may have a typical length of the order of several tens of micro-meters)subjected to the given loading condition would most probably grow, then you should be justified to claim that the seal may not survive. Otherwise you may of course claim the opposite.
Well, that procedure works well only if the assumptions made concerning the distribution of crack-length of pre-existing flaws and other imperfections in the bulk rubber is rather good.
But even if you do not know ( ...as is usually the case)the mentioned "distribution" you can get benefit out of the above described "calculus" when you for example state that you have in your rubber component a rather big number of small cracks with a crack-length of 50 micro-meters or so.
Despite of these limitations I would suggest the following procedure to follow in order to receive a prognoses for service life of the o-ring in question:
Perform a non-linear finite element analysis of your sealing problem, where a typical cyclic loading should be simulated. Detect the areas of the cross section where strains are large. Get the full information on the strain tensor in those areas. Determine the components of the strain tensor which are in tension, because they may be the most dangerous ones. If you have the impression that one of the tension-strain components is much greater than the other ones (which often appears to be)you may be justified to handle that deformation mode for further considerations only. Now correlate the strain energy density found in the considered area of the cross section with the tear energy appropriate for the given deformation mode (see for example "How to design rubber components" by A. Gent). Thus you will get a relation dealing with a virtual crack you think about to be present in the considered area of the cross section of the o-ring. Then you may assume you would know the critical tearing energy ( ... and indeed one can assume a "most-lower" value - to be on the safe side). With all those assumption in mind you can state whether the current dynamical loading may be critical concerning a possible growth of some virtual cracks (flaws etc of certain crack-length). If you thus come to the conclusion that initial cracks ( always be present, independently of compound quality which may have a typical length of the order of several tens of micro-meters)subjected to the given loading condition would most probably grow, then you should be justified to claim that the seal may not survive. Otherwise you may of course claim the opposite.
Well, that procedure works well only if the assumptions made concerning the distribution of crack-length of pre-existing flaws and other imperfections in the bulk rubber is rather good.
But even if you do not know ( ...as is usually the case)the mentioned "distribution" you can get benefit out of the above described "calculus" when you for example state that you have in your rubber component a rather big number of small cracks with a crack-length of 50 micro-meters or so.
Thanks ParaDoc. I will look into your suggestions. One other condition that may affect at least part of my analysis is the size of the unsupported space. If there is enough of a cavity for the seal to extrude into, then I may be experiencing some yielding so to speak and the phenomenon may not be entirely fatigue. Thanks again for all the help.
...well, you mean probably a kind of extrusion of the seal into a gap or the like? In that case the stain may be extraordinary large, but yielding will hardly occur, because you should keep in mind that you are dealing with a rubbery material. But what could take place is the following: the tearing energy associated with this strain could surmount a critical value and therefore initiate crack growth of probably very small, always existing, initial cracks (flaws etc). So under cyclic loading of the seal part of the material, which is repeatedly deformed into the mentioned gap will disintegrate and heavy wear will take place ...
- Status