Fix bug while processing ACP's imported solid models (#469)

* Fix bug while processing analysis plies of imported solid models.

* update unit test

---------

Co-authored-by: Kathy Pippert <[email protected]>

examples/010_harmonic_example.py

@@ -50,7 +50,7 @@
 import matplotlib.pyplot as plt
 
 from ansys.dpf.composites.composite_model import CompositeModel
-from ansys.dpf.composites.constants import FailureOutput
+from ansys.dpf.composites.constants import FAILURE_LABEL, FailureOutput
 from ansys.dpf.composites.example_helper import get_continuous_fiber_example_files
 from ansys.dpf.composites.failure_criteria import (
     CombinedFailureCriterion,
@@ -119,7 +119,6 @@
 # This is a bit confusing, but DPF uses the same label for Frequency and Time
 FREQ_LABEL = "time"
 PHASE_LABEL = "phase"
-FAILURE_LABEL = "failure_label"
 all_phases_and_freqs_failure_value_fc = dpf.FieldsContainer()
 all_phases_and_freqs_failure_value_fc.labels = [FREQ_LABEL, PHASE_LABEL]
 

src/ansys/dpf/composites/layup_info/_layup_info.py

@@ -324,14 +324,30 @@ def get_analysis_ply_index_to_name_map(
     ----------
     mesh
         DPF Meshed region enriched with lay-up information
+
+    .. note::
+
+        Analysis plies of ACP's imported solid model that are linked only
+        to homogeneous elements are currently skipped.
     """
     analysis_ply_name_to_index_map = {}
     with mesh.property_field("layer_to_analysis_ply").as_local_field() as local_field:
+        element_ids = local_field.scoping.ids
         for analysis_ply_name in get_all_analysis_ply_names(mesh):
             analysis_ply_property_field = _get_analysis_ply(
                 mesh, analysis_ply_name, skip_check=True
             )
             first_element_id = analysis_ply_property_field.scoping.id(0)
+            # analysis plies which represent a filler ply are ignored because
+            # they are linked to homogeneous elements only. So, they are not
+            # part of layer_to_analysis_ply. This filler plies can occur in
+            # imported solid models.
+            # The analysis ply indices can be retrieved from
+            # analysis_ply_property_field as soon as the
+            # properties of PropertyField are available in Python.
+            if first_element_id not in element_ids:
+                continue
+
             analysis_ply_indices: list[int] = local_field.get_entity_data_by_id(first_element_id)
 
             layer_index = analysis_ply_property_field.get_entity_data(0)[0]

src/ansys/dpf/composites/server_helpers/_versions.py

@@ -53,7 +53,7 @@ class _DpfVersionInfo:
     "9.0": _DpfVersionInfo(
         "9.0",
         "2025 R1 pre 0",
-        "DPF Composites: exposure of ply type.",
+        "DPF Composites: exposure of ply type and support of imported solid models.",
     ),
 }
 

tests/composite_model_scoping_test.py

@@ -26,7 +26,7 @@
 import pytest
 
 from ansys.dpf.composites.composite_model import CompositeModel, CompositeScope
-from ansys.dpf.composites.constants import FailureOutput
+from ansys.dpf.composites.constants import FAILURE_LABEL, FailureOutput
 from ansys.dpf.composites.data_sources import get_composite_files_from_workbench_result_folder
 from ansys.dpf.composites.failure_criteria import (
     CombinedFailureCriterion,
@@ -35,7 +35,7 @@
     MaxStressCriterion,
 )
 from ansys.dpf.composites.layup_info import get_all_analysis_ply_names
-from ansys.dpf.composites.server_helpers._versions import version_older_than
+from ansys.dpf.composites.server_helpers._versions import version_equal_or_later, version_older_than
 
 from .helper import get_basic_shell_files
 
@@ -49,13 +49,13 @@ def test_composite_model_element_scope(dpf_server, data_files):
 
     composite_scope = CompositeScope(elements=[1, 3])
     failure_container = composite_model.evaluate_failure_criteria(cfc, composite_scope)
-    irfs = failure_container.get_field({"failure_label": FailureOutput.FAILURE_VALUE})
+    irfs = failure_container.get_field({FAILURE_LABEL: FailureOutput.FAILURE_VALUE})
     min_id = irfs.scoping.ids[np.argmin(irfs.data)]
     max_id = irfs.scoping.ids[np.argmax(irfs.data)]
 
     composite_scope = CompositeScope(elements=[min_id, max_id])
     max_container = composite_model.evaluate_failure_criteria(cfc, composite_scope)
-    max_irfs = max_container.get_field({"failure_label": FailureOutput.FAILURE_VALUE})
+    max_irfs = max_container.get_field({FAILURE_LABEL: FailureOutput.FAILURE_VALUE})
     assert len(max_irfs.data) == 2
     assert max_irfs.get_entity_data_by_id(min_id)[0] == pytest.approx(min(irfs.data), 1e-8)
     assert max_irfs.get_entity_data_by_id(max_id)[0] == pytest.approx(max(irfs.data), 1e-8)
@@ -78,7 +78,7 @@ def test_composite_model_named_selection_scope(dpf_server, data_files, distribut
 
     scope = CompositeScope(named_selections=[ns_name])
     failure_container = composite_model.evaluate_failure_criteria(cfc, scope)
-    irfs = failure_container.get_field({"failure_label": FailureOutput.FAILURE_VALUE})
+    irfs = failure_container.get_field({FAILURE_LABEL: FailureOutput.FAILURE_VALUE})
     assert len(irfs.data) == 2
     assert irfs.get_entity_data_by_id(2) == pytest.approx(1.4792790331384016, 1e-8)
     assert irfs.get_entity_data_by_id(3) == pytest.approx(1.3673715033617213, 1e-8)
@@ -118,8 +118,8 @@ def test_composite_model_ply_scope(dpf_server):
 
     scope = CompositeScope(plies=ply_ids)
     failure_container = composite_model.evaluate_failure_criteria(cfc, scope)
-    irfs = failure_container.get_field({"failure_label": FailureOutput.FAILURE_VALUE})
-    modes = failure_container.get_field({"failure_label": FailureOutput.FAILURE_MODE})
+    irfs = failure_container.get_field({FAILURE_LABEL: FailureOutput.FAILURE_VALUE})
+    modes = failure_container.get_field({FAILURE_LABEL: FailureOutput.FAILURE_MODE})
 
     if version_older_than(dpf_server, "7.0"):
         # the old implementation did not allow to distinguish between plies of the different parts.
@@ -194,8 +194,8 @@ def test_composite_model_named_selection_and_ply_scope(dpf_server, data_files, d
 
     scope = CompositeScope(named_selections=[ns_name], plies=ply_ids)
     failure_container = composite_model.evaluate_failure_criteria(cfc, scope)
-    irfs = failure_container.get_field({"failure_label": FailureOutput.FAILURE_VALUE})
-    modes = failure_container.get_field({"failure_label": FailureOutput.FAILURE_MODE})
+    irfs = failure_container.get_field({FAILURE_LABEL: FailureOutput.FAILURE_VALUE})
+    modes = failure_container.get_field({FAILURE_LABEL: FailureOutput.FAILURE_MODE})
     assert len(irfs.data) == 2
     assert len(modes.data) == 2
     expected_irfs_by_id = {2: 0.49282684, 3: 0.32568454}
@@ -230,6 +230,53 @@ def test_composite_model_time_scope(dpf_server):
     for time, expected_max_irf in time_id_and_expected_max_irf.items():
         scope = CompositeScope(time=time)
         failure_container = composite_model.evaluate_failure_criteria(cfc, scope)
-        irfs = failure_container.get_field({"failure_label": FailureOutput.FAILURE_VALUE})
+        irfs = failure_container.get_field({FAILURE_LABEL: FailureOutput.FAILURE_VALUE})
         assert len(irfs.data) == 4
         assert max(irfs.data) == pytest.approx(expected_max_irf, abs=1e-6)
+
+
+def test_ply_wise_scoping_in_assembly_with_imported_solid_model(dpf_server):
+    """Ensure that the ply-wise scoping works in combination with the reference surface plot."""
+    if version_older_than(dpf_server, "8.0"):
+        pytest.xfail("Not supported because of limitations in the handling of assemblies.")
+
+    result_folder = pathlib.Path(__file__).parent / "data" / "assembly_imported_solid_model"
+    composite_files = get_composite_files_from_workbench_result_folder(result_folder)
+
+    # Create a composite model
+    composite_model = CompositeModel(composite_files, dpf_server)
+
+    plies = [
+        "Setup 2_shell::P1L1__ModelingPly.2",
+        "Setup_solid::P1L1__ModelingPly.2",
+        "Setup 3_solid::P1L1__ModelingPly.1",
+    ]
+
+    # Evaluate combined failure criterion
+    combined_failure_criterion = CombinedFailureCriterion(failure_criteria=[MaxStressCriterion()])
+    failure_result = composite_model.evaluate_failure_criteria(
+        combined_criterion=combined_failure_criterion, composite_scope=CompositeScope(plies=plies)
+    )
+
+    # check the on reference surface data
+    for failure_output in [
+        FailureOutput.FAILURE_VALUE_REF_SURFACE,
+        FailureOutput.FAILURE_MODE_REF_SURFACE,
+        FailureOutput.MAX_GLOBAL_LAYER_IN_STACK,
+        FailureOutput.MAX_LOCAL_LAYER_IN_ELEMENT,
+        FailureOutput.MAX_SOLID_ELEMENT_ID,
+    ]:
+        field = failure_result.get_field({FAILURE_LABEL: failure_output})
+
+        if version_equal_or_later(dpf_server, "9.0"):
+            assert field.size == 21
+        elif version_equal_or_later(dpf_server, "8.0"):
+            # Servers of the 2024 R2 series do not extract the reference surface
+            # of imported solid models. So the reference surface mesh
+            # contains only the shell elements and reference surface of the
+            # standard solid model.
+            assert field.size == 12
+        else:
+            # Servers before 8.0 are not tested because of several limitations:
+            # handling of assemblies, reference surface not supported
+            assert False

tests/composite_model_test.py

@@ -40,6 +40,7 @@
 from ansys.dpf.composites.layup_info import (
     LayerProperty,
     LayupModelContextType,
+    get_all_analysis_ply_names,
     get_analysis_ply_index_to_name_map,
 )
 from ansys.dpf.composites.layup_info.material_properties import MaterialMetadata, MaterialProperty
@@ -539,3 +540,72 @@ def test_failure_criteria_evaluation_default_unit_system(dpf_server):
     cfc = CombinedFailureCriterion("max stress", failure_criteria=[MaxStressCriterion()])
 
     failure_container = composite_model.evaluate_failure_criteria(cfc)
+
+
+def test_composite_model_with_imported_solid_model_assembly(dpf_server):
+    """
+    Tests the show on reference surface option for imported solid models.
+
+    Imported solid models are slightly different if compared with shell parts
+    and standard solid models. For instance, they have analysis plies which are
+    not linked to a layered elements. These are called filler plies.
+
+    In addition, the `show on reference surface` option uses the skin of the
+    imported solid mesh instead of a predefined reference surface.
+    This general test ensures that the composite model can handle these.
+
+    The model has one shell part, one standard solid model and
+    an imported solid model.
+    """
+    result_folder = pathlib.Path(__file__).parent / "data" / "assembly_imported_solid_model"
+    composite_files = get_composite_files_from_workbench_result_folder(result_folder)
+
+    # Create a composite model
+    composite_model = CompositeModel(composite_files, dpf_server)
+
+    # ensure that all analysis plies are available, also the filler plies
+    # The initial version of DPF Composites had limitations in the handling
+    # of assemblies and so the test is skipped for old versions of the server.
+    if version_equal_or_later(dpf_server, "7.0"):
+        analysis_plies = get_all_analysis_ply_names(composite_model.get_mesh())
+        ref_plies = [
+            "Setup 2_shell::P1L1__ModelingPly.2",
+            "Setup_solid::P1L1__ModelingPly.2",
+            "Setup 3_solid::P1L1__ModelingPly.1",
+            "Setup 3_solid::P1L1__ModelingPly.3",
+            "Setup 3_solid::filler_Epoxy Carbon Woven (230 GPa) Prepreg",
+            "Setup 3_solid::P1L1__ModelingPly.2",
+            "Setup 2_shell::P1L1__ModelingPly.1",
+            "Setup_solid::P1L1__ModelingPly.1",
+        ]
+        assert set(analysis_plies) == set(ref_plies)
+
+    # Evaluate combined failure criterion
+    combined_failure_criterion = CombinedFailureCriterion(failure_criteria=[MaxStressCriterion()])
+    failure_result = composite_model.evaluate_failure_criteria(combined_failure_criterion)
+
+    irf_field = failure_result.get_field({FAILURE_LABEL: FailureOutput.FAILURE_VALUE})
+    assert irf_field.size == 84
+
+    # check the on reference surface data
+    for failure_output in [
+        FailureOutput.FAILURE_VALUE_REF_SURFACE,
+        FailureOutput.FAILURE_MODE_REF_SURFACE,
+        FailureOutput.MAX_GLOBAL_LAYER_IN_STACK,
+        FailureOutput.MAX_LOCAL_LAYER_IN_ELEMENT,
+        FailureOutput.MAX_SOLID_ELEMENT_ID,
+    ]:
+        field = failure_result.get_field({FAILURE_LABEL: failure_output})
+        if version_equal_or_later(dpf_server, "9.0"):
+            assert field.size == 60
+        elif version_equal_or_later(dpf_server, "8.0"):
+            # Servers of the 2024 R2 series do not extract the reference surface
+            # of imported solid models. So the reference surface mesh
+            # contains only the shell elements and reference surface of the
+            # standard solid model.
+            assert field.size == 18
+        else:
+            # Server of 2024 R1 and before do not support results on the reference
+            # surface at all. It is tested that the operator update
+            # completes without error nevertheless.
+            pass