How to access shell information with DPF.

Victor Trejo

Hi! I would like to know how to query the following shell layer information with DPF (C++):

• Thickness
• Orientation angle
• Number of integration points
• Material ID

Thanks in advance!
Victor Trejo.

Mike.Thompson

Seems like you want this operator:
https://dpf.docs.pyansys.com/version/stable/operator_reference_load.html

op = dpf.operators.result.mapdl_section() # operator instantiation
op.inputs.properties_name.connect(my_properties_name)
op.inputs.section.connect(my_section)# optional
op.inputs.streams_container.connect(my_streams_container)
op.inputs.data_sources.connect(my_data_sources)
op.inputs.layer_property.connect(my_layer_property)# optional
op.inputs.layers_requested.connect(my_layers_requested)# optional
my_properties_value = op.outputs.properties_value()
my_layers_per_section = op.outputs.layers_per_section()

Ayush Kumar

@Victor Trejo can you please confirm which version of DPF Server are you using?
layer_property and its results are only supported from 2024R1 which will be released next month.
For 2023R2 you can only extract Thickness, NumLayers.

Victor Trejo

Thank you very much, Mike!

That indeed looks like what I am looking for. Would you happen to have the equivalent in C++? Alternatively, could you please point me to a similar workflow in C++? I am confused about some of the input pins, particularly about the streams_container.

Mike.Thompson

I don’t believe you need to use the streams container if you point to a data source like a result file.

Victor Trejo

You are right. I tried without the streams container and it worked correctly. Thank you very much, Mike!

Victor Trejo

I have the impression that the layer_property pin is being ignored. I am getting the expected data for Thickness and NumLayers, but I am getting all zeroes for Orientation and NumIntPoints.

Also, when I try to get the layers_per_section properties field by calling mapdl.getOutputPropertyField(1), I get the following runtime error:
"DPF error - runtime error: self_call; requested result not found".

Am I misusing DPF?

This is my code snippet:

ansys::dpf::Operator mapdl("mapdl_section_properties");
mapdl.connect(4, dataSources);
mapdl.connect(5, true); // layer property
mapdl.connect(0, std::vector<std::string>{"Thickness", "NumLayers", "Orientation", "NumIntPoints"});
auto fieldContainer = mapdl.getOutputFieldsContainer(0);
auto size = fieldContainer.size();

std::cout << "thickness\n";
auto thickness = fieldContainer.at(0);
std::cout << thickness.describe() << '\n';

std::cout << "numLayers\n";
auto numLayers = fieldContainer.at(1);
std::cout << numLayers.describe() << '\n';

std::cout << "orientation\n";
auto orientation = fieldContainer.at(2);
std::cout << orientation.describe() << '\n';

std::cout << "integrationPointsCount\n";
auto integrationPointsCount = fieldContainer.at(3);
std::cout << integrationPointsCount.describe() << '\n';

std::cout << "layersPerSection\n";
mapdl.run();
auto layersPerSection = mapdl.getOutputPropertyField(1);
std::cout << layersPerSection.describe() << '\n';

Victor Trejo

Hi! I was wondering if anyone had any insight into my last question. It would be really appreciated since this is currently blocking my progress.

Thanks in advance!
Victor.

Mike.Thompson

@Victor Trejo if you think there is a legitimate defect in the code, please file it in the appropriate GitHub system.
I don’t have a direct solution for you currently, but one of the thing to look into would be PyMAPDL where you could also retrieve this data in a different way through more traditional APDL methods.

Victor Trejo

Thank you very much for your answer, @Mike.Thompson!

Which is the appropriate GitHub system for DPF (C++, not PyDPF)?

Using PyMAPDL is not an option for the project I am working on, it needs to be the C++ DPF library for legal reasons.

Victor Trejo

Hi @Ayush Kumar,

Thank you very much for the information. We are using DPF 20230723.1, it looks like I have to wait until next month. Could you please share the release date?

Ayush Kumar

Hi @Victor Trejo,
sorry I don't have any official date yet. Please check customer portal for further information.