Overview
This is a tool for generating ABAQUS input data for the metal plasticity and damage material models. It can:
- Process uniaxial tensile coupon data to generate ABAQUS input for the nonlinear plastic parameters.
- Generate ABAQUS input for the nonlinear plastic parameters from user-defined material parameters.
- Generate ABAQUS input for the nonlinear plastic parameters from auto-generated material parameters, based on a large material database for European (and equivalent) carbon steel grades S235, S275 and S355, and high-strength bolt grades 8.8 and 10.9.
- Generate ABAQUS input for the ductile damage parameters of bolt steel grades 8.8 and 10.9. This input is mesh-independent.
Using the tool
- Input: choose Manual (pick a steel grade, or type your own values) or Test data (browse for a coupon file, paste two columns, or pick one of the two sample coupons, carbon steel or stainless steel, from Examples).
- Randomize samples E, σy, σu, εu and εf from normal distributions fitted to the database; use Resample for a new draw.
- Plot updates as you type. Hover or tap to read values; save it as PNG, SVG or CSV.
- ABAQUS input: the Plastic table is the true stress vs plastic true strain for *Plastic. Generate plastic data saves it as Mat-True-Stress-Strain-Plastic-for-ABAQUS-Input.txt; the True σ–ε tab saves Mat-True-Stress-Strain.txt. Copy table copies tab-separated values that paste straight into Excel or the ABAQUS/CAE data table.
- Damage parameters (bolt grades): set the mesh size Lmesh and characteristic length Lc, then save the damage initiation table as Mat-Bolt_damage.txt and use uf for damage evolution.
Manual stress–strain curves
The engineering curve follows the two-stage model of Yun & Gardner (2017) for hot-rolled carbon steel, with strain-hardening onset εsh = 1.05 σy/E:
σ = σy + (σu − σy) [0.4 α + 2α / (1 + 400 α5)1/5], α = (ε − εsh)/(εu − εsh)
True values are εt = ln(1 + ε) and σt = σ(1 + ε) up to necking. Beyond necking the lower-bound power law (Ling 1996) is used up to εt,f = ln(1 + εf):
σt = σu(1 + εu) (εt / εt,u)εt,u
Plastic strain is measured from the yield point, and the first row of the plastic table is set to σy,0.2. The dashed red branch is an indicative post-necking engineering curve through (εu, σu) down to 0.85 σy,0.2 at εf; it is not written to ABAQUS.
Coupon test data
- The file holds engineering strain (absolute values, not %) and engineering stress in MPa. Files over 1000 rows are thinned to about 1000.
- Negative strains are removed, the first point is zeroed, strain reversals are dropped, and before the peak only points with rising stress are kept.
- σu and εu come from the peak; εf is the last strain. σy is the first point where the tangent slope falls between 0 and 10 000 MPa; E is a linear fit to the points before it. σy,0.2 uses the 0.2% offset line.
- The true curve runs up to necking, then one more point is added at ln(1 + εf) using half the necking-zone slope from the last six points.
Material database
Mean (and standard deviation used with randomize):
| Grade | E [MPa] | σy [MPa] | σu/σy | εu | εf |
|---|
| S235 | 197482 (10346) | 301 (45) | 1.51 (0.162) | 0.15 (0.03) | 0.31 (0.06) |
| S275 | 206468 (10228) | 324 (29) | 1.49 (0.162) | 0.18 (0.03) | 0.30 (0.05) |
| S355 | 205674 (7405) | 395 (28) | 1.35 (0.10) | 0.15 (0.01) | 0.28 (0.04) |
| 8.8 | 214265 (6976) | 786 (96) | 1.17 (0.10) | 0.078e−0.010 L, 0.271e−0.011 L |
| 10.9 | 211267 (10255) | 1003 (65) | 1.13 (0.07) | 0.070e−0.011 L, 0.274e−0.019 L |
Without randomize the S-grade σu means are 440, 477 and 529 MPa, and 902 and 1133 MPa for the bolts. Sampled values are bounded below by the nominal grade values (e.g. σy ≥ 355, σu ≥ 470 MPa for S355) and E ≥ 180 000 MPa. Bolt strains depend on the mesh size L = Lmesh [mm] (randomize adds σ = 0.003 on εu and 0.012 / 0.014 on εf).
Ductile damage (bolts)
With the plastic strain at necking εu,pl = εu − σu/E, C1 = εu,pl/4 and C2 = εu,pl, the equivalent plastic strain at damage initiation versus stress triaxiality η is:
η ≤ −1/3: 10000 · −1/3 < η ≤ 0: C1/(1 + 3η) · 0 < η ≤ 1/3: C1 + (C2 − C1)(3η)2 · η > 1/3: C2/(3η)
The table covers η = −0.5 … 1.0 in steps of 0.01 with strain rate 0. The displacement at failure for damage evolution is
uf = (εf − εu) Lc
where Lc is the element characteristic length (press ? beside Lc for guidance). Lc cannot be smaller than Lmesh.
ABAQUS keywords
*Elastic
E, 0.3
*Plastic
<Plastic table: yield stress, plastic strain>
*Damage Initiation, criterion=DUCTILE
<Damage table: fracture strain, triaxiality, strain rate>
*Damage Evolution, type=DISPLACEMENT
uf
Citation
Elkady, A. (2023) “ABAQUS_SteelMat_Generator: A tool for generating metal plastic and damage parameters for ABAQUS”, Zenodo, Version v23.04. doi:10.5281/zenodo.7756886
Ding, Z. and Elkady, A. (2025) “Generalized spring model for steel bolts in tension considering uncertainty and loading speed”, Journal of Constructional Steel Research, 231, 109574. doi:10.1016/j.jcsr.2025.109574
Author: Dr Ahmed Elkady, University of Southampton
To report bugs or give feedback, raise an issue through the GitHub repository or e-mail a.elkady@soton.ac.uk.
Other references
- Yun, X. and Gardner, L. (2017). Stress-strain curves for hot-rolled steels. J. Constructional Steel Research, 133, 36–46.
- Ling, Y. (1996). Uniaxial true stress-strain after necking. AMP Journal of Technology, 5, 37–48.
Web version of ABAQUS Steel Generator v26-01. Relative to the desktop app: the plot and tables update live, the true curve is always shown, raw coupon data is drawn behind the processed curve, uf and the damage table use the displayed εu, εf, and the (inactive) cyclic-hardening options are not shown.
Disclaimer
These tools are provided for research and educational purposes only, without warranty of any kind. The generated material data are estimates from the cited models and your inputs, and are only as reliable as those models within their validity ranges. Results should be checked independently by a qualified engineer before being used in design or assessment.
