FIT THE HARDENING CURVES AND RETURN PARAMETERS¶
Fits a hardening curve (effective strain vs. effective stress) to a selected analytical model — Swift, Voce, Hockett-Sherby, or Stoughton — using a neural-network optimisation loop. Returns the fitted model parameters as a dataset. Use this worker when you need calibrated hardening law coefficients for material card authoring or downstream forming/crash simulations.
When to use¶
Classification: process.
Tagged: curve_fit, hardening, hockett, material_parameters, neural_network, plasticity, stoughton, stress_strain.
Inputs¶
| Label | ID | Type | Default | Required | Description |
|---|---|---|---|---|---|
| HC-ESVES | hardening_curve-effective_strainvs_effective_stress | vector | — | Input hardening curve as a two-column vector of (effective plastic strain, effective stress) data points (dimensionless strain, stress in MPa or model-consistent units); leave empty only if the curve is supplied upstream in the workflow. | |
| Type Of Fit | typeof_fit | list | swift | Hardening law to fit: ‘swift’ (power-law, default), ‘voce’ (saturating exponential), ‘hockett’ (Hockett-Sherby), or ‘stoughton’; choose based on the material’s strain-hardening character. | |
| I-HTB | iterations-higherthebetter | list | 10 | Maximum number of neural-network optimisation iterations (1–1000); higher values improve fit accuracy at the cost of runtime — default 10 is suitable for quick checks, use 250–1000 for production calibrations. | |
| LR-LTB | learning_rate-lowerthebetter | list | 0.1 | Gradient-descent learning rate for the neural-network optimiser (1 down to 1e-5); smaller values yield more stable convergence — default 0.1 is a reasonable starting point, reduce to 0.001 or below for noisy or stiff curves. | |
| T-LTB | tolerance-lowerthebetter | list | 1e-05 | Tolerance - Lower the better | |
| NOE-POP-LTF | numberof_evaluations-percentageof_points-lowerthefaster | list | 10 | Number of Evaluations - Percentage of Points - Lower the faster | |
| Number Of Digitization Points | numberof_digitization_points | list | 100 | Number of Digitization Points | |
| Return Type | return_type | list | curve |
Outputs¶
| Label | ID | Type | Description |
|---|---|---|---|
| curve_fit_hardening_neuralnetwork_parameters_output_1 | curve_fit_hardening_neuralnetwork_parameters_output_1 | dataset | Dataset containing the fitted hardening-law parameters (e.g. C, ε₀, n for Swift; R₀, R∞, b for Voce) together with goodness-of-fit metrics, ready for downstream material card authoring or further post-processing. |
Disciplines¶
- ai_ml.supervised.regression
- engineering.material.calibration
- engineering.material.characterization
Runnable example¶
A runnable example is registered for this worker. Open the example workflow on the d3VIEW canvas: /api/workflow/example?id=curve_fit_hardening_neuralnetwork_parameters
Auto-generated from transformation schema. Worker id: curve_fit_hardening_neuralnetwork_parameters. Schema hash: 6bc1611ca1d2. Hand-curated docs in workerexamples/ override this page when present.