.. _auto_curve_fit_hardening_neuralnetwork:

*FIT THE HARDENING CURVES AND RETURN THE CURVE*
===============================================

Fits a strain-hardening curve (effective strain vs. effective stress) to a user-selected analytical model (Swift, Voce, Hockett, Stoughton, or MHS) using a neural-network optimiser. Returns the fitted curve for direct use in material calibration workflows.

When to use
-----------

Classification: **process**.

Tagged: ``curve_fit``, ``hardening``, ``hockett``, ``mat_24``, ``material_calibration``, ``mhs``, ``neural_network``, ``stoughton``.

Inputs
------

.. list-table::
   :header-rows: 1
   :widths: 20 20 20 20 20 20

   * - Label
     - ID
     - Type
     - Default
     - Required
     - Description
   * - HC-ESVES
     - hardening_curve-effective_strainvs_effective_stress
     - vector
     - —
     - 
     - Input strain-hardening curve as an (effective strain, effective stress) vector pair (dimensionless vs. MPa or consistent stress units); the curve to be fitted — optional if a default test dataset is preloaded.
   * - Type Of Fit
     - typeof_fit
     - list
     - swift
     - 
     - Hardening law to fit: 'swift' (power-law), 'voce' (saturation), 'hockett' (Hockett-Sherby), 'stoughton', or 'mhs' (mixed hardening); defaults to 'swift'.
   * - Iterations
     - iterations-higherthebetter
     - list
     - 10
     - 
     - Maximum number of neural-network training iterations (1–1000); higher values improve fit accuracy at the cost of compute time — default of 10 is suitable for quick checks, use 250–1000 for final calibration.
   * - LR-LTB
     - learning_rate-lowerthebetter
     - list
     - 0.1
     - 
     - Gradient-descent learning rate for the neural-network optimiser (1 down to 1e-5); lower values yield more stable convergence — default of 0.1 is a reasonable starting point, reduce to 0.001 or below if the loss oscillates.
   * - T-LTB
     - tolerance-lowerthebetter
     - list
     - 1e-05
     - 
     - Lower the better
   * - Number Of Evaluations
     - numberof_evaluations-percentageof_points-lowerthefaster
     - list
     - 10
     - 
     - 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
-------

.. list-table::
   :header-rows: 1
   :widths: 20 20 20 20

   * - Label
     - ID
     - Type
     - Description
   * - curve_fit_hardening_neuralnetwork_output_1
     - curve_fit_hardening_neuralnetwork_output_1
     - vector
     - Fitted hardening curve as an (effective strain, effective stress) vector evaluated with the optimised analytical-model parameters; ready for downstream MAT_24 keyword authoring or further material-calibration workers.

Disciplines
-----------

- ai_ml.supervised.regression
- data.curve.transform
- 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 <https://www.d3view.com/api/workflow/example?id=curve_fit_hardening_neuralnetwork>`_

.. raw:: html

   <hr style="margin-top:2em">
   <p style="font-size:11px;color:#888">
   Auto-generated from <code>transformation</code> schema. Worker id: <code>curve_fit_hardening_neuralnetwork</code>. Schema hash: <code>74d1e26a3669</code>. Hand-curated docs in <code>workerexamples/</code> override this page when present.
   </p>