FM-73 Adhesive Mech Properties

[VN1981]

Hi Folks,
I am trying to model a 2D layer of FM-73 structural adhesive using FE. I was able to locate some of the mech properties for the adhesive but they are mainly in-plane ones. I am guessing in order to determine peel stresses, I require out-of-plane ones as well. Can I use G12=G23=G13?

 

[VN1981]

I found complete set of Mechanical properties for FM 73 adhesive at this paper:

Link:

http://www3.nd.edu/~kmatous/Papers/COMPB_TGJ.pdf

Title: Analysis of tongue and groove joints for thick laminates

Authors: Karel Matous, George J. Dvorak, Department of Mechanical, Aerospace, and Nuclear Engineering, Centre for Composite Materials and Structures, Rensselaer Polytechnic Institute

Here are the properties.

E11=E22=E33 = 403204 psi (2.778 GPa)
G12=G13=G23 = 145907 psi (1.006 Gpa)
ν12= ν23= ν13 = 0.38
Fult = 7.25 ksi (50 MPa)

Another ref:

Title: Elastic–Plastic Analysis and Strength Evaluation of Adhesive Joints in Wind Turbine Blades

Authors: Yi Hua, Ananth Ram, Linxia Gu, Mech & Materials Engineering, University of Nebraska

Longitudinal Modulus, E1 = 159,542 psi (1.1 GPa)
In-Plane Shear Modulus, G12 = 55,405 psi (0.382 GPa)
Major In-plane Poisson's Ratio ν12 = 0.44

Some of the properties differ by quite a bit. Which one looks closer to what is accepted by industry for FM 73 adhesive?

 

[blakmax]

Whoa! Adhesive properties change dramatically with temperature. You would need to be sure that the properties you are using are correct for the case being addressed. A further consideration is that average shear stress is MEANINGLESS in bonded joints, so quoting Fult means nothing. An average shear stress model predicts that doubling the overlap will result in twice the load being carried and that is simply not true. A further issue is that adhesives are elastic-plastic with MOST of the strain energy to failure coming from plastic behaviour, so unless your model is elastic-plastic you will not get reliable strength predictions. It is possible to calculate load at failure for a bonded joint using the elastic-plastic methods from Hart-Smith. The important thing here is that you are calculating LOAD not stress.

I can help with elastic plastic design data if you want to go down that path.

Regards

Blakmax

 

[ESPcomposites]

Here are some sources (most public domain) and a few have FM 73 properties.

Note that adhesive properties are "effective" properties that are a function of test method, adhesive thickness, and other factors. They are also a function of environment. Lastly, properties related to shear strength are reasonably useful but attempting to use peel stresses in a useful manner may be a lost cause for a practical solution (unless you have a lot of joint test data to make sense of it). You usually want to reduce to peel stresses to insignificance as it relates to the bondline. Finally, if you are using this for a joint with composite adherends, then you may have to address the peel mode of the adherend (interlaminar failure). This mode is usually critical because the adhesive is usually tougher (more plasticity) than the resin system (more brittle). That problem is better address with interlaminar fracture mechanics (as opposed to classical nonlinear stress analysis of the adhesive bondline).

---

Military Handbook MIL-HDBK-17-3F, Composite Materials Handbook, Vol. 3, Chapter 6, June, 2002.

Tomblin, J., Senevirante, W., Escobar, P., and Yoon-Khian, “Shear Stress-Strain Data for Structural Adhesives,” DOT/FAA/AR-02/97, November, 2002.

Tomblin, C., Yang, C., Harter, P., “Investigation of Thick Bondline Adhesive Joints,” DOT/FAA/AR-01/33, June, 2001.

Engineering Sciences Data Unit, ESDU 81022, “Guide to the Use of Data Items in the Design of Bonded Joints.”

Tan, S.C., “Analysis of Bolted and Bonded Composite Joints,” WR-TR-92-4084, Interim Report, September 1992.

Tomblin, J., Seneviratne, W., Escobar, P., Yap, Y., “Fatigue and Stress Relaxation of Adhesives in Bonded Joints,” DOT/FAA/AR-03/56, October, 2003.




Brian

http://www.espcomposites.com

 

[VN1981]

Blakmax & ESP, thanks for your replies.

ESP, I hear your comments. I've never worked with bonded joints before. All my previous experience has been with metallic structures and some bolted composites. The exercise I am trying to do is to get a handle on the methodology of calculating check for various failure modes from FE for a bonded joint Thanks for providing me with the various public domain info. I will start with MIL-17. Totally forgot that resource could contain info about FM-73.

> Finally, if you are using this for a joint with composite adherends, then you may have to address the peel mode of the adherend (interlaminar failure). This mode is usually critical because the adhesive is usually tougher (more plasticity) than the resin system (more brittle).
.One of the adherend is composite & the other is Aluminum. I am trying to reproduce an investigation which was conducted by Sandia National Lab project. Thanks for the insight on adherend peel mode failure. I had read about it but it completely slipped off my mind.

> That problem is better address with interlaminar fracture mechanics (as opposed to classical nonlinear stress analysis of the adhesive bondline).

If you have time, can you please point me to good & lucidly illustrated/explained resources on how to do the above procedure. I am interested in learning.

Thanks again for the reply.

 

[ESPcomposites]

There are a lot of papers about interlaminar fracture mechanics (ILFM) and it is a fairly involved topic. I think Krueuger has some pretty well written ones that give an overview.

Krueger, R. “The Virtual Crack Closure Technique: History, Approach and Applications,” NASA/CR-2002-211628, April, 2002.

Krueger, R., “Virtual Crack Closure Technique: History, Approach, and Applications,” Applied Mechanics Reviews, Vol. 57, No. 2, March 2004, pp. 109-143.

CMH-17 has some information as well.

My book (due late this year hopefully) will also cover both bondline analysis via stress analysis and interlaminar failure modes via ILFM.


Brian

http://www.espcomposites.com

 

[VN1981]

Thanks again ESP.

I am aware of your upcoming book and it is something I am eagerly awaiting.

I have to ask for one more help. Could you point out any material which describes detailed or semi-detailed procedure or methodology of FE modeling of adhesively bonded joints? I am particularly interested in learning how to do various joint checks from FE results.

I am assuming much of the material will be proprietary but I am hoping some of it may be from academic/DOT/FAA sources and could be open source (?)!

Thanks for the sources on ILFM. I will go through it when I have some more time.

Have a good weekend!

 

[ESPcomposites]

Well, that is like opening a can of worms. I can't completely answer that here, but can give some brief thoughts:

- For a joint with metal adherends, the analysis is concerned with the bondline. Lets start with the cases where the peel stresses are reduced to insignificance via design (tapered edges, joint stabilization, etc.). The simplest (and maybe most reliable approach) is to use classical solutions or a computational approach that give you just the average through-thickness shear stresses (and validate with test). The thick adherend test is commonly used to generate material properties. You can also use a FEM, but there are several issues related to how to model and properly extract the relevant stresses to be compared apparent material properties. I won't be able to go into detail here though since that can be involved, but it is not terribly difficult. You just have to realize that you are after average through-thickness bondline stresses and be aware that the FEM can generate some extra information and capture additional stresses.

- Next, consider a joint with metal adherends where the peels stresses are significant enough to affect the joint load capability. There aren't really any great classical solutions for this and FEM is often used to determine the peel stresses. But the major issues is what to do with these peel stresses that the FEM has provided. They are highly localized and have steep gradients. Furthermore, while there are tests to determine the apparent through-thickness strength of adhesives, I don't believe there is a proven way to correlate any apparent material properties with these highly localized stresses. Finally, a nonlinear elastic-plastic approach would be needed and there is no established interaction equation for shear/peel of adhesives in the nonlinear range. Some use V-M up to yield, but that won't help you too much. So anything you find related to this is likely to just going to be a topic of research interest. In reality, you want to design to minimize the peel stresses to insignificance (validated by test). Short of that, you will have to do joint tests to capture this effect. Some then create their own blend of how to correlate this to an analysis (such as a point stress approach) or you just go to a solely empirical approach.

- For composites, you have to address the delamination failure mode and it will likely be critical (as opposed to the bondline). This mode is inherently sensitive to the interlaminar tensile stresses (peel stresses). A fracture mechanics solution is more appropriate to address these localized high stress gradient areas and the somewhat brittle nature of typical resin systems. There are ASTM standardized tests for Mode I and Mode II (recently) which can be used to correlate to the FEM. There are few fracture mechanics based approaches and VCCT may be the most popular. That is an inherently FEM based solution so any papers on that will automatically discuss FEM. Note that this is still a topic of research, but is gaining acceptance. Joint tests will still be needed to validate the approach.

Brian

http://www.espcomposites.com

 

[Compositepro]

If you are going to do any FEA on bonded joints you should start with the shear stress-strain curves from the KGR-1 test instrument.

 

[blakmax]

Here is the data extracted using ASTM D3983 which is based on the KGR extensometer. The reference is enclosed, or you may use the RAAF publication AAP 7021.016-1. Note that the data has been terminated at a shear strain of 1.0. The data is for 0.06psf FM73. If you want 0.08psf data I can supply that as well.

Regards

Blakmax

 

[VN1981]

ESP, Blakmax & Compositepro,
Thanks for your replies again. Mucho appreciated.

Since my interest is personal self-development, I have no access to test equipment. But I do agree with the process i.e. FE results have to be correlated with real world testing. I guess, I will have to do with the best I can with just FE software and hopefully, some of you folks can guide me along the way to learn the methodology.

 

[VN1981]

I began my modeling a simple repair scenario. Aluminum 2024-T3 skin with a damage repaired using a bonded composite patch. The adhesive layer is FM-73 and the doubler is Boron-Epoxy laminate.

The Al Skin is of 0.080" thick and the doubler is 0.075" thick.

I learnt from Cytec official publications that the nominal thicknesses of FM-73 adhesive are 0.007", 0.010" & 0.015". I made an assumption that just one layer of adhesive of thickness 0.010" is used in the repair. The adhesive layer was modeled as Isotropic.

All the 3 components were modeled as 2D shell elements in Patran. The loading case was simple tension. Currently, I am restricting the analysis to simple linear elastic and I may conduct a elastic-plastic as my learning progresses.

When I check the results in the adhesive layer for out-of-plane shear stresses, I am getting the value as zero. I think that by modeling as 2D, a plane stress condition is assumed and hence the stress through the thickness will be constant/zero.

So my questions are as follows:

1. To get through the thickness variation of stresses in the adhesive, is a 3D FE formulation necessary?

2. I was hoping that due to eccentricity between the base skin & repair, there would be some minor out of plane loading in adhesive layer and this would get reflected in the out of plane shear values. Am I incorrect with respect to point 2?

I would appreciate some suggestions based on your experiences folks.

Thanks in advance...