Material Characterization

[kenaycock]

Hi Karthik,

Thank you for collecting the material characterization data. Here are the most recent results from my fit of the ABAQUS UMAT to the data you provided:

and here are the UMAT material constants:

Parameter	Value
Ea	59,000 MPa
𝜈a	0.33
Em	24,750 MPa
𝜈m	0.33
εL	0.042
𝛿σ/𝛿TL	4.77 MPa / °C
σSL	395 MPa
σEL	435 MPa
T0	37 °C
𝛿σ/𝛿TU	9.00 MPa / °C
σSU	165 MPa
σEU	145 MPa
σScL	395 MPa
εLV	0.042
NA	0
NP	3
σP1	1205 MPa
εP1	0.090
σP2	1350 MPa
εP2	0.100
σP3	1450 MPa
εP3	0.120

The ABAQUS stress-strain data were extracted from a single C3D8R element loaded in uniaxial tension. The logarithmic strain (LE22) and the Cauchy stress (S22) from the element integration point are displayed in the plot.

A couple of questions:

1. Since compression tests were not performed, I set the material constant for the start of solid-solid phase transformation in compression (σ^S_cL) to be the same as that for tension (σ^S_L). However, you mentioned that σ^S_cL can be important for accurately predicting the behavior of specimens loaded in bending. Do you recommend that we use a scaling factor, e.g., σ^S_cL = 1.25 σ^S_L?
2. You also mentioned the idea of performing material characterization using a geometry that mimics that of the specimen used in the experimental tensile tests (i.e., a round tube). However, in my verification tests, I simply used a single cubic element. Should we repeat the verification using a round tube geometry?

I uploaded my analysis files here if you’d like to take a closer look:
https://github.com/kenaycock/Generic-IVC-Filter/tree/master/Material-Characterization

Thanks,
Kenny

[karthikSenthi]

Hi Kenny,

To answer your questions.

1. We normally assume a scaling factor of σScL = 1.30 σSL. When you perform validation studies like the surrogate fatigue testing and if the force vs. displacement does not fit well, then we can try considering different scaling factors.

2. I misunderstood your original question of the fit being over stiff in the austenite-martensite transformation regime. I thought you already performed validation studies like the surrogate testing and the fit was over stiff compared to the validation testing. In those cases when the validation is not good, we need to make sure the geometry is representative of the actual test geometry. The single element material fit that you performed is good. We do not need to repeat it on a round tube, as our final geometry will be some form of laser cut representation. Coming to your original question of material fit being over stiff in the austentite-martensite transition region compared to the tensile tests, that is because of the presence of R-phase in the material. As you can see there are 2 distinct slopes, one for Austenite and one for R-phase before it goes into the upper plateau. The R-phase is an intermediate martensite phase that exists in some cases depending on the processing history. The code does not have the functionality to capture the R-phase and so the way you have defined it is good since we only need the Austenite modulus.

The material parameter you have documented in the Table is good and you could change the start of the compression loading based on a 1.3 scaling factor. If you could also include a picture of the full strain so that we see the plasticity points of the data it will be helpful for others.

Thanks,
Karthik

[kenaycock]

Hi Karthik,

After reviewing the V&V standards more closely, we think it would be a good idea to acquire some compression data for our material to ensure that we are separating our "validation" and "calibration" activities (e.g., see ASME V&V-10, section 3.4.1). We wouldn't need the full hysteresis data for compression, just the material constant for the start of solid-solid phase transformation in compression (σ^S_cL).

Could you look into this for us and let us know if this parameter could be measured in future material characterization data?

Thanks,
Kenny