Add anisotropic material mechanics by streeve · Pull Request #264 · ORNL/CabanaPD

streeve · 2025-11-06T19:57:27Z

Still missing actual dependence on orientation; most of the models still default to isotropic only (all of LPS, etc.)

src/force/CabanaPD_PMB.hpp

pabloseleson · 2025-12-10T19:53:46Z

src/CabanaPD_ForceModels.hpp


 template <typename PeridynamicsModelType, typename MechanicsModelType = Elastic,
-          typename DamageType = Fracture,
+          typename AnisotropyType = Isotropic, typename DamageType = Fracture,


Why we use here "AnisotropyType = Isotropic"?

The default anisotropy is "no anisotropy"

src/force_models/CabanaPD_LPS.hpp

pabloseleson · 2025-12-10T20:03:20Z

unit_test/tstForce.hpp

                    // matter which particle we check.
-                    f_mag = model( CabanaPD::ForceCoeffTag{}, 0, 0, s0, vol );
+                    f_mag = model( CabanaPD::ForceCoeffTag{}, 0, 0, s0, vol,
+                                   0.0, xi, xi_x, xi_y, xi_z );


Is "0.0" input argument arbitrary?

Should be removed now

pabloseleson · 2025-12-10T20:05:12Z

src/force_models/CabanaPD_PMB.hpp

    using base_type::delta;
    using base_type::K;
-    double c;
+    double _c;


Could you please remind me what is this change for?

Variable and (new) function can't have the same name

pabloseleson · 2025-12-10T20:10:07Z

src/force_models/CabanaPD_PMB.hpp

+    BaseForceModelPMB( PMB model, NoFracture fracture, Cubic,
+                       const double delta, const double _C11,
+                       const double _C12 )
+        : base_type( model, fracture, delta, 1.0 / 3.0 * ( _C11 + 2.0 * _C12 ) )


What is "1.0 / 3.0 * ( _C11 + 2.0 * _C12 )" for?

We need K to create the base model. The previous assumption was that every model uses K

…ticity

pabloseleson · 2026-03-16T16:13:25Z

src/force/CabanaPD_LPS.hpp


                // Break if beyond critical stretch unless in no-fail zone.
-                if ( model( CriticalStretchTag{}, i, j, r, xi ) &&
+                if ( model( CriticalStretchTag{}, i, j, r, xi, xi_z ) &&


Note adding xi_z only (instead of all the bond components) is specific for transversely isotropic. Other symmetries would require xi_x and/or xi_y.

pabloseleson · 2026-03-16T16:14:14Z

src/force/CabanaPD_PMB.hpp


                // Break if beyond critical stretch unless in no-fail zone.
-                if ( model( CriticalStretchTag{}, i, j, r, xi ) &&
+                if ( model( CriticalStretchTag{}, i, j, r, xi, xi_z ) &&


Same comment as above.

pabloseleson · 2026-03-16T16:26:25Z

src/force_models/CabanaPD_PMB.hpp

+        , C33( _C33 )
+    {
+        A1111 = 75.0 / 4.0 * C11 - 75.0 / 2.0 * C13 + 15.0 / 4.0 * C33;
+        A1133 = -25.0 / 3.0 * C11 + 115.0 / 2.0 * C13 + 5.0 / 4.0 * C33;


This should be:

A1133 = -25.0 / 3.0 * C11 + 115.0 / 2.0 * C13 - 15.0 / 2.0 * C33;

pabloseleson · 2026-03-16T16:27:40Z

src/force_models/CabanaPD_PMB.hpp

+    BaseForceModelPMB( PMB model, Elastic, NoFracture fracture,
+                       TransverselyIsotropic, const double delta,
+                       const double _C11, const double _C13, const double _C33 )
+        : base_type( model, fracture, delta, 1.0 / 3.0 * ( _C11 + 2.0 * _C13 ) )


What is "1.0 / 3.0 * ( _C11 + 2.0 * _C13 )" for?

pabloseleson · 2026-03-16T16:42:05Z

src/force_models/CabanaPD_PMB.hpp

+    }
+
+    KOKKOS_FUNCTION
+    auto lambda( const double xi, const double xi1, const double xi2,


We may need to think about whether to use xi_x, xi_y, xi_z, instead of xi1, xi2, xi3 for consistency with the rest of the code.

pabloseleson · 2026-03-16T16:47:14Z

src/force_models/CabanaPD_PMB.hpp

+                     const double xi_y, const double xi_z,
+                     const int = -1 ) const
+    {
+        return lambda( xi, xi_x, xi_y, xi_z ) * s * vol;


This function should be "c" instead of \lambda.

Note:

and

So, c = \lambda * xi^3

pabloseleson · 2026-03-16T16:52:02Z

src/force_models/CabanaPD_PMB.hpp

+                     const double xi_y, const double xi_z,
+                     const int = -1 ) const
+    {
+        return lambda( xi, xi_x, xi_y, xi_z ) * s * vol;


Same comment as above (c vs. \lambda)

pabloseleson · 2026-03-16T16:52:29Z

src/force_models/CabanaPD_PMB.hpp

+    KOKKOS_INLINE_FUNCTION
+    auto operator()( EnergyTag, const int, const int, const double,
+                     const double, const double, const int = -1 ) const
+    {


We need to implement this, also for transversely isotropic.

pabloseleson · 2026-03-16T16:56:37Z

src/force_models/CabanaPD_PMB.hpp

        // 0.25 factor is due to 1/2 from outside the integral and 1/2 from
        // the integrand (pairwise potential).
-        return 0.25 * c * stretch_term * xi * vol;
+        return 0.25 * c() * stretch_term * xi * vol;


What is this for? Bind energy has stretch^2

pabloseleson · 2026-03-16T16:57:54Z

src/force_models/CabanaPD_PMB.hpp

+        , s_Y( sigma_y )
+    {
+        for ( std::size_t d = 0; d < s_Y.size(); d++ )
+            s_Y[d] /= ( 3.0 * base_type::K );


This is not anisotropic since it depends on K and not K[d]

pabloseleson · 2026-03-16T16:59:53Z

src/force_models/CabanaPD_PMB.hpp

+
+        // Must extract again if in the plastic regime.
+        s_p = _s_p( i, n );
+        return lambda( xi, xi_x, xi_y, xi_z ) * ( s - s_p ) * vol;


Same comment c vs. \lambda aboive

pabloseleson · 2026-03-16T17:00:02Z

src/force_models/CabanaPD_PMB.hpp

+    auto operator()( EnergyTag, const int, const int, const double,
+                     const double, const double, const int = -1 ) const
+    {
+        // TODO: implement energy.


pabloseleson · 2026-03-16T17:03:39Z

src/CabanaPD_ForceModels.hpp

+                     const double xi, const double xi_z ) const
    {
+        auto bond_break_coeff =
+            s0[0] + ( s0[1] - s0[0] ) * xi_z * xi_z / ( xi * xi );


This is wrong. For isotropic:

bond_break_coeff = ( 1.0 + s0 ) * ( 1.0 + s0 );

streeve added 4 commits November 6, 2025 09:54

WIP

e0c7256

Add initial infrastructure for anisotropic materials

9cf55c7

fixup: simplify anisotropic model

f4b4ec6

FIXME: this is where the physics is needed

40f3d3c

streeve requested a review from pabloseleson November 6, 2025 19:57

streeve self-assigned this Nov 6, 2025

streeve added 3 commits November 7, 2025 15:17

Add cubic anisotropy for PMB

e937323

fixup: device markup

42260fb

fixup: tag naming

412f2f7

pabloseleson reviewed Nov 25, 2025

View reviewed changes

src/force/CabanaPD_PMB.hpp Outdated Show resolved Hide resolved

streeve added 2 commits December 9, 2025 12:12

fixup: cubic compile

8e8cb2b

DO NOT MERGE: change elastic wave to anisotropic

f92f749

pabloseleson reviewed Dec 10, 2025

View reviewed changes

src/force_models/CabanaPD_LPS.hpp Show resolved Hide resolved

pabloseleson reviewed Dec 10, 2025

View reviewed changes

pabloseleson and others added 13 commits December 10, 2025 15:30

Revised cubic model

16db62f

Incorporated back unintended deletion of operator

fc0db66

fixup: simplify constructor

cd6d481

fixup: remove r in aniso

2605b06

fixup: separate isotropic/cubic base models

e483ab3

Added transversely isotropic model

fe31311

fixup: compile transverse iso

e2753ab

fixup: delegating constructor

a91146b

Initial EPP with TransverseIso

3a78460

DO NOT MERGE: use tensile test as starting point for anisotropic plas…

2faa78e

…ticity

Implemented initial orientation-dependent yield stretch

59790e2

fixup: compile

fe45761

First draft anisotropic bond breaking

b7c4034

fixup: wrong s0 for aniso

837b87c

streeve mentioned this pull request Dec 30, 2025

Physics roadmap #112

Open

13 tasks

fixup: actually pass two values for aniso dogbone

9f531bb

pabloseleson reviewed Mar 16, 2026

View reviewed changes

streeve and others added 2 commits March 16, 2026 10:25

Fix incorrect transvse isotropic bond breaking criterion

4189a94

Fixed an anisotropic elastic constant and bond breaking

3c1467c

Conversation

streeve commented Nov 6, 2025

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

pabloseleson Mar 16, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

pabloseleson Mar 16, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

pabloseleson Mar 16, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

pabloseleson Mar 16, 2026 •

edited

Loading

pabloseleson Mar 16, 2026 •

edited

Loading

pabloseleson Mar 16, 2026 •

edited

Loading