About the *.eld file and Chamis model (Modified RoM).

[UOMO]

Dear Louise,

I read your post:
http://texgen.sourceforge.net/phpBB3/vi ... amis#p6227

I understand that the intention is to provide at an element-level, properties of the yarn based on local volume fractions.
For this, you make use of the semi-empirical modified rule of mixtures provided by C.C. Chamis. (Figure attached taken from the NASA TM https://ntrs.nasa.gov/archive/nasa/casi ... 008360.pdf ).

Therefore, one needs (at an element level) the value of the local fiber volume fraction. My questions are:

1. I can see that the *.eld file have the same number of elements. What is what is given in each column or where the information of the local volume fraction is provided?
2. If the *.eld file contains the information of the local volume fraction (point 1. is true), how does TexGen calculate this (algorithm)?

Any text or reference about the algorithm that TexGen uses to calculate the Local volume fractions in the yarns, would be highly appreciated.

Best!.

[louisepb]

Hi Uomo,

If you do a search for .eld in the .inp file you will find a description of the entries in the .eld file. (I should replicate this at the top of the .eld file really).
Information on volume fraction calculation can be found in section 2.10 and Appendix A of Martin Sherburn's thesis. When generating the .eld file TexGen automatically calculates the yarn cross-section at the plane of the element cross-section and uses this to calculate the local area. You will need to have input the yarn properties in order for correct results to be calculated.
There are classes which describe the fibre distribution. At the moment TexGen uses the constant distribution assuming that the fibres are distributed evenly across the yarn.

Hope that helps,
Louise

[UOMO]

Let

Hi Uomo,

If you do a search for .eld in the .inp file you will find a description of the entries in the .eld file. (I should replicate this at the top of the .eld file really).
Information on volume fraction calculation can be found in section 2.10 and Appendix A of Martin Sherburn's thesis. When generating the .eld file TexGen automatically calculates the yarn cross-section at the plane of the element cross-section and uses this to calculate the local area. You will need to have input the yarn properties in order for correct results to be calculated.
There are classes which describe the fibre distribution. At the moment TexGen uses the constant distribution assuming that the fibres are distributed evenly across the yarn.

Hope that helps,
Louise

Dear Louise,
let me see if I understood correctly. I am using TexGen 3.10.0. I am then exporting inp/ori/eld files using the volume mesh option. If I search for ".eld" in the .inp file, it cannot find anything.

1. Anyways, I am trying to figure out how the .eld file works. It is a file with a number of rows equal to the number of elements and 6 columns. The first column is the index of the element. The second row seems to be a -1 for the matrix and then if four yarns, it has numbers from 0 to 3. Then, there are four columns left, what do they mean? I could say than from cols. 3-5, it is a coordinate, but the last one are negative numbers, so, I guess it is not a volume fraction, right? None of the columns seems to have fully positive numbers. Or, are they vectors?.
I can see from the FORTRAN file chamis_model_final.for that is says: C STATEV(4) - LOCAL FIBRE VOLUME FRACTION (READ FROM *.ELD FILE). But still I cannot figure out from where.

2. Also, I looked at Martin Sherburn's thesis, appendix A where it is shown the calculation of an area fraction of fibre to yarn. Moreover, in Eq. (2.68) one can find the "linear density" term, just right before defining 2.68. Is it already assumed a certain density for the fibers before calculating the volume fractions in the .eld file?. or, is it used instead Eq. (2.70) and then it is assumed constant?. By using Eq. (2.70), I understand that the integral(A.F., dx) along the length should be equal to the total V.F. of fiber in a yarn, right? If this is the case, which "dx" is it assumed to calculate the local volume fraction in every cross-section? (Or am I confusing something?).

I hope you could explain a bit more on this... .

Thank you a lot and sorry for the long questions!,
Best,
/UOMO.

[louisepb]

Hi Uomo,

Apologies, I apparently only added the header to the .eld file in the voxel export. (Something else for my to-do list!) It's here:
********************
*** ELEMENT DATA ***
********************
** Element data stored as a depvars
** 1 - Yarn Index (-1 for matrix, first yarn starting at 0)
** 2/3 - Location (x and y cross-section coordinates of element relative to yarn centerline)
** 4 - Volume fraction
** 5 - Distance of element from the surface of the yarn (for yarn elements only, distance is negative)

The linear density is a yarn property and will be given on the yarn data sheet. Given the linear density and the fibre density it is possible to calculate the area of fibres across the section of the yarn. It is assumed by default that the fibres are distributed linearly across the section (although it would be possible to change this). TexGen calculates the area of the yarn for the given cross=section and from this and the fibre area calculates the local volume fraction. Because the fibre distribution across the section is linear the volume fractions across a section will be the same, but if the cross section varies along the length of the yarn then the Vf will vary along the length.

I hope that helps,
Louise

[UOMO]

Hi Uomo,

Apologies, I apparently only added the header to the .eld file in the voxel export. (Something else for my to-do list!) It's here:
********************
*** ELEMENT DATA ***
********************
** Element data stored as a depvars
** 1 - Yarn Index (-1 for matrix, first yarn starting at 0)
** 2/3 - Location (x and y cross-section coordinates of element relative to yarn centerline)
** 4 - Volume fraction
** 5 - Distance of element from the surface of the yarn (for yarn elements only, distance is negative)

The linear density is a yarn property and will be given on the yarn data sheet. Given the linear density and the fibre density it is possible to calculate the area of fibres across the section of the yarn. It is assumed by default that the fibres are distributed linearly across the section (although it would be possible to change this). TexGen calculates the area of the yarn for the given cross=section and from this and the fibre area calculates the local volume fraction. Because the fibre distribution across the section is linear the volume fractions across a section will be the same, but if the cross section varies along the length of the yarn then the Vf will vary along the length.

I hope that helps,
Louise

Dear Louise,

thank you for the explanation. So, to conclude, if I understood correctly, one gets the area of the yarns (A.Y., at a cross section), the area of the fibers (A.F., at a cross section using the linear density) and then the ratio A.F./A.Y. is taken the same as the volume fraction V.F./V.Y at a given cross section?. Or am I missing something?.

On the other hand, if I take the sum of the local volume fractions in a yarn, say yarn0, then sum(index_of_elements_in_yarn0,4) should be something close to 100? Or not necessarily? I will double check this column 4, I recall having negative values.
Thanks again!.

[louisepb]

Hi,
Yes that's correct.

I don't think you can generate anything meaningful by summing the local volume fractions. There will be theoretical maximum value for the Vf depending on how the fibres are packed but this will never be 100 - that would be a solid yarn.

Best wishes,
Louise

[UOMO]

Hi,
Yes that's correct.

I don't think you can generate anything meaningful by summing the local volume fractions. There will be theoretical maximum value for the Vf depending on how the fibres are packed but this will never be 100 - that would be a solid yarn.

Best wishes,
Louise

Excellent Louise,
thank you a lot for clarifying this. Now everything makes sense.
Best!
Juan José.