While conduction a simulation using the Pocan B 3215 material (PBT with 10% glass fibers) I noticed that apparently one can choose a micro-mechanics model without having neither filler nor matrix data.
Moldflow offers a handful of models (like the Halpin-Tsai, Tandon-Weng, ect..) for calculating the mechanical properties. Normally all these models require data for the fiber as well as for the matrix. But when I take a look at the material properties in the database, there is data neither for the filler nor for the matrix. Instead there is anisotropic ‘mechanical properties data’.
Am I right assuming that the composite´s anisotropic properties already were calculated by the manufacturer? If that is so, why can I still select a model under the ‘Process Settings’ tab?
Such a model is supposed to give out the composite properties but without filler data it seems useless to me…
So I wonder how does Moldflow consider these models in the simulation?
Thanks in advance for your replies.
Solved! Go to Solution.
Well, in fact the Moldflow database has the mecanical data for the fibers:
For the composite mechanical properties and for the matrix properties. To see the later you have to change the matrix properties menu from "do not use matrix properties" to "Use 7 matrix properties" (or to the obsolete use 6 matrix properties):
The "composite properties" measured by manufacturer or by Moldflow lab a characteristics of the composite for one particular fiber orientation (usually A11~0.85, A22~0.1, A33~0.05 but it can vary significantly depending on how the sample was manufactured). Moldflow uses micro-mechanical models to predict composite properties in every point of the moldings. They vary quite significantly from point to point and those variations are usually the major contributing factor to the warpage. So even if the composite properties are measured (for one particular orientation state) we have to calculate the properties for all other orientation states. We use the micromechanical models for this. The fiber orientation is predicted by Moldflow, the filler properties are in the database. There are two methods for getting the matrix properties depending on the "Use matrix" drop box in the "Mechanical properties" tab:
- If we select "Use 7 matrix properties" or "Use 6 matrix properties" then the matrix properties from the database are used.
- If we select ("Do not use matrix properties") then the software decomposes the matrix properties and the testing sample orientation from the composite properties and filler properties. This way the composite properties for the testing sample would not change with the micromechanical model (although the variation of the properties certainly will). For isotropic matrices it is usually the default option, while for the anisotropic matrices the default is "Use 7 matrix properties"
Thanks a lot, for responding so quickly.
Now, do I get this right: when the ‘composite properties’ were measured (by manufacturer or by Moldflow) the program calculates back the properties of the matrix and uses them for the micro-mechanics model? So e.g. the values of E1=5000 MPa and E2=3400 Mpa are measured from that sample (A11~0.85, A22~0.1, A33~0.05) and the matrix’ E1 and E2 (2968.99 MPa) are calculated by Moldflow?
Will in this case the choice of a different micro-mechanics model affect my simulation results?
I am using Moldflow for my master thesis and for some of the material´s properties I have Data that was obtained in the university’s department. Precisely I want to use also a user defined ‘coefficient of thermal expansion’. Do I have to use the mean CTE value or the derivative CTE?
Yes, you get it right. By default Moldflow software starts from the composite properties and calculates matrix properties. The choice of the micromechanical model is still important though - the "composite properties" that we match are just for one fiber orientation state, for other fiber orientations the result will depend upon the micro-mechanical model.
Moldflow assumes that LCTE is for the room temperature (it adjusts the value based on PVT data). Thus, you need to enter the differential CTE at T=20C (I guess the mean differential CTE for the range e.g.. 0..40C would be good as well). Measurements of CTE at elevated temperatures may include various visco-elastic effects, hysteresis, etc - that is why we prefer not to rely on them