Newbie: elastomer modelling for pump. Looking for input

[cnuk]

I am tasked with becoming familiar with FEA analysis of elastomers. We built motors (see http://www.pvfluid.com/prodserv.html for animated example) out of nitrile rubber that undergo relatively small strains (.02" to .06" compression on a thickness of approx .5"), but sometimes at a high rate (100+Hz). This of course leads to heat buildup and the failure of our product. Some of the things I would like to be able to analyze are: contact pressure between the two mating parts as a means of calculating seal integrity; shear stress at the bond surface where elastomer is bonded to steel outer tube; friction between the two components, particularly as the elastomer part warms up and thermally expands; and temperature due to hysteresis. This is my wish list, but I am trying to be realistic. I currently model some of these things with a linear FEA package by only inputting thermal exp coeff, elastic modulus and poissons ratio. Now that you've stopped laughing at my simplistic approach I must say that's all I can do with the tools I have. Another division of our company purchased a DMTA a year or so ago and we are now getting what appears to be very detailed material properties (elastic and viscous modulus, tan delta, thermal exp coeff, etc. ). I am trying to justify a new FEA package to take advantage of this information and improve our product. I could ramble for hours, but a few of my questions are:
1. How would you rate this problem in terms of complexity? I am a mechanical engineer not an analyst. I would like to learn how to do the analysis but some people still tell me this is only for full time gurus. Ours is a very parametric product line and we would always be solving very similar problems.
2. If you had to pick a FEA package which would it be? I see lots of reference to Abaqus but have also read lots about MSC Marc and Ansys being more than capable of solving my problem. I don't want to pay for capabilities I will never use.
3. Any good advice on getting started in the exciting new world of elastomer analysis? Good books? Good websites(other than this one)?

Thank You very much.

[Jorgen]

Hello Cnuk,

You are asking important questions, and I am sure there are many other people in a similar situation.

All the things that you want to analyze can done. Some of them are easier than others. The most difficult is probably the thermal heat up, since that will require a coupled mechanical and thermal analysis.

The DMTA is a great tool for characterizing the materials that you are working on. Based on my experience, I would also claim that you can gain a lot of insight into your products, how they behave and how to examine and optimize different designs, by using FEA.

About your specific questions:
1. The complicity is not too bad. There will be a learning curve for the FE software, but the actual simulations should be rather routine.

2. The three major packages that are used today are: ABAQUS, Ansys, and MSC Marc. All three of these software packages are very capable and well suited for your needs. There are other packages that are less expensive that might also work for you, e.g. ls-dyna, adina, etc. In my oppinion, the three major FE packages mentioned above tend to have better documentation and support, which are very important in an industrical setting.

3. There are a few books that might be of interest to you. One of the better ones is: "Science and technology of Rubber" by James E. Mark. There are not many high quality web-sites about elastomer analysis - that's why I created this site :wink:

Best of luck,
Jorgen

[cnuk]

Thank you for the quick, and informative, response. I was like a kid at christmas when I found your website today. Lots of good information to digest. Now I have to work on selling my boss on the idea. He wants to consult everything out from now until the end of time ($$$), yet I think it should be a core competency and I should become proficient. Time to work on selling FEA, which is harder than it should be. Some people don't believe...

Once again, thank you!