Bug located in Setup add_mesh_link

[SanderBovee]

Before submitting the issue

• I have searched among the existing issues
• I am using a Python virtual environment

Description of the bug

For my maxwell transient simulation i'm trying to link the mesh to my static solution.
In PYAEDT 0.10.0 not passing a solution argument was finding the correct solution name however now in PYAEDT 0.15 i'm noticing that wrong solution name is being used and passing a solution as argument yields the wrong result.

Steps To Reproduce

1. Create maxwell magnetostatic design -> setup/solve
2. Create a maxwell transient design in the same project.
   A setup.add_mesh_link("Bfield",solution= "MySetupAuto : LastAdaptive") yields a mesh link to "default"
   A setup.add_mesh_link("Bfield") yields a mesh link to "MySetupAuto : Transient"

Which Operating System are you using?

Windows

Which Python version are you using?

3.10

Installed packages

annotated-types==0.7.0
ansys-api-edb==1.0.10
ansys-edb-core==0.1.9
ansys-pythonnet==3.1.0rc6
anyio==4.9.0
app==0.1.0
attrs==24.2.0
cffi==1.17.1
click==8.1.8
clr_loader==0.2.7.post0
colorama==0.4.6
defusedxml==0.7.1
exceptiongroup==1.2.2
fastapi==0.115.12
fonttools==4.56.0
fpdf2==2.8.2
grpcio==1.70.0
h11==0.14.0
idna==3.10
Jinja2==3.0.3
jsonschema==4.23.0
jsonschema-specifications==2024.10.1
MarkupSafe==3.0.2
msgpack==1.0.8
nest-asyncio==1.6.0
numpy==1.26.4
occ-utils==1.2.4
occ-wrapper==7.3.0.1
packaging==24.2
pandas==2.2.3
parapy==1.13.1
parapy-lib-construction==0.3.3
parapy-webgui==1.2.6
parapy-webgui-layout==0.15
parapy-webgui-mui==1.1.2
parapy-webgui-plotly==0.8.0
parapy-webgui-projects==0.10.7
parapy-webgui-reflex==0.9.0
parapy-webgui-viewer==0.25.7
pillow==11.1.0
plumbum==1.9.0
protobuf==4.25.6
psutil==7.0.0
pyaedt==0.15.5
pycparser==2.22
pydantic==2.10.6
pydantic_core==2.27.2
pyedb==0.39.1
python-dateutil==2.9.0.post0
python-multipart==0.0.20
pytomlpp==1.0.13
pytz==2025.2
pywin32==310
PyYAML==6.0.2
referencing==0.35.1
rpds-py==0.24.0
rpyc==6.0.1
rtree==1.4.0
scikit-rf==1.6.2
scipy==1.15.2
six==1.17.0
sniffio==1.3.1
starlette==0.46.1
toml==0.10.2
tomli==2.2.1
tomli_w==1.2.0
typing_extensions==4.13.0
tzdata==2025.2
uvicorn==0.34.0
watchdog==3.0.0
websockets==15.0.1
wxPython==4.2.1

[anur7]

Hi @SanderBovee. Are you importing the mesh from Maxwell magnetostatic (source design) to Maxwell transient? Have you tried to look at the mesh if the target design and the source design have the same mesh? The results could be wrong because the mesh is not the same for both source and target design

[SanderBovee]

Hi @anur7 , Thanks for looking into it.

I'm trying to reuse the magnetostatic mesh for the transient design. Maxwell magnetostatic can optimize its mesh based on "problem" area's. But for transient designs no such feature exists. Which is understandable considering the simulation time required if every timestep also needs to optimize its own mesh.

But this means that the transient model doesn't have mesh. The only requirement is that the model in both designs needs to be identical. Which i can confirm if i just manually add the mesh link in the electronics desktop UI. Which then runs perfectly.

I'm just unable to provide the correct name to the solution it needs to link to and the automatic selection of the solution name isn't working correctly either.

I've also verified this by logging the ANSYS actions to a script:
Running setup.add_mesh_link("Bfield")
Yields:
oModule.EditSetup("MySetupAuto",
[
"NAME:MySetupAuto",
"Enabled:=" , True,
[
"NAME:MeshLink",
"ImportMesh:=" , True,
"Project:=" , "This Project*",
"Product:=" , "Maxwell",
"Design:=" , "Bfield",
"Soln:=" , "MySetupAuto : Transient",
[
"NAME:Params"
],
"ForceSourceToSolve:=" , True,
"PreservePartnerSoln:=" , True,
"PathRelativeTo:=" , "TargetProject",
"ApplyMeshOp:=" , True
],
"NonlinearSolverResidual:=", "0.0001",
"ScalarPotential:=" , "Second Order",
"SmoothBHCurve:=" , False,
"StopTime:=" , "100ms",
"TimeStep:=" , "pwl(TimeDataSet,time)",
"OutputError:=" , False,
"OutputPerObjectCoreLoss:=", True,
"OutputPerObjectSolidLoss:=", False,
"UseControlProgram:=" , False,
"ControlProgramName:=" , "",
"ControlProgramArg:=" , "",
"CallCtrlProgAfterLastStep:=", False,
"FastReachSteadyState:=", False,
"AutoDetectSteadyState:=", False,
"IsGeneralTransient:=" , True,
"IsHalfPeriodicTransient:=", False,
"SaveFieldsType:=" , "None",
"UseNonLinearIterNum:=" , False,
"CacheSaveKind:=" , "Count",
"NumberSolveSteps:=" , 1,
"RangeStart:=" , "0s",
"RangeEnd:=" , "0.1s"
])
But soln should be "MySetupAuto : LastAdaptive". If i click in the ansys UI on "Edit link" ansys instantly jumps to the correct name aswell. (from " Transient" to "LastAdaptive")

[anur7]

@SanderBovee thank you for your clarification. From what I understood, if I pass the correct name for "solution" argument: setup.add_mesh_link("Bfield",solution= "MySetupAuto : LastAdaptive"), it works. Instead, if I don't specify the "solution": setup.add_mesh_link("Bfield"), it links to "MySetupAuto : Transient" which is not correct. Please correct me if I'm wrong

[SanderBovee]

@anur7 Hi Anur, partly correct. if you dont specify a solution the automatic selection is wrong. But if you provide a solution name it points to something else entirely?

In my example i specified [solution= "MySetupAuto : LastAdaptive"] but when i looked in ANSYS i saw that the mesh link was made to "default". But where this is coming from i have no idea.

[gmalinve]

Hi @SanderBovee,

sorry for the delay but I was trying to figure out what was causing the issue.
I tested both versions (0.10 and latest) and I found bugs in both actually.
The reason why it works in 0.10 is just out of luck. We hardcoded the solution to always be last adaptive (in your case worked because you are importing the mesh from Magnetostatic where the nominal adaptive is "your_setup_name : LastAdaptive"). This is wrong as we know because for transient is "your_Setup_name : Transient". Please look at the implementation:

elif (
                solution.split()[0] in self.p_app.oproject.GetDesign(design).GetChildObject("Analysis").GetChildNames()
            ):
                meshlinks["Soln"] = "{} : LastAdaptive".format(solution.split()[0])

The other bug I found in latest version (and the reason why you see the "Default" solution) is because we don't set the source solution when it is passed as an argument in the method. If you are interested please take a look at my fix in this branch: https://github.com/ansys/pyaedt/pull/6133/files.

This fix will be available in the next release.

Thanks for reporting this to us!

Giulia

[SanderBovee]

hi @gmalinve,
Thank you !