isciml: generate-mesh tdem-loop
Last updated
Last updated
The generate-mesh tdem-loop subcommand in isciml is used to create simulation-ready mesh structures with distinct zones, customizable dimensions, and adjustable refinement levels for time-domain electromagnetic induction loop based data acquisition. This command is essential as the first step in synthetic data generation, producing the base mesh upon which simulations are run to generate structured training and test data. It allows for fine-grained control over mesh characteristics, tailored to the requirements of complex synthetic data simulations.
The general syntax for using commands is:
--inner-loop <FLOAT>
Sets the dimensions (x, y, z) of the inner loop zone.
(1.0, 1.0, 1.0)
No
--modeling-zone <FLOAT>
Sets the dimensions (x, y, z) of the modeling zone.
(12.0, 12.0, 12.0)
No
--outer-box <FLOAT>
Sets the dimensions (x, y, z) of the outer box.
(50000.0, 50000.0, 50000.0)
No
--plane-z
Z-coordinate for the reference plane.
0.0
No
--loop-z <FLOAT>
Set the vertical location of the loop
0.0
No
--element-size-inner <FLOAT>
Element size within the inner loop.
0.25
No
--element-size-modeling <FLOAT>
Element size within the modeling zone.
0.5
No
--element-size-outer <FLOAT>
Element size within the outer box.
6250.0
No
--output-folder <TEXT>
The folder to store the generated Salome Python script and output mesh.
"output_folder"
No
--script-name <TEXT>
Name of the generated Python script.
"generated_mesh_script.py"
No
--template <TEXT>
The path to the Salome template script (e.g., mesh_template.py).
"mesh_template.py"
No
The generate-mesh tdemp-loop subcommand is designed to create a multi-zone mesh structure with customizable zone dimensions and refinement levels. This mesh serves as the foundational structure for generating simulation data, enabling users to define precise characteristics for inner loops, modelling zones, and outer boxes. Each zone can be assigned unique element sizes, allowing for highly accurate simulation models. Additionally, the generate-mesh tdemp-loop command supports template-based script generation, making it easy to replicate complex mesh configurations tailored to different physics requirements.
This command is essential for setting up the physical model needed for synthetic data generation. Once the mesh is created, simulations can be run on this structure to produce data suitable for AI model training, test sets, or advanced analytics.
To create a mesh with default output names and specific dimensions:
The generated mesh includes multiple zones, each defined with its own dimension and element size, ensuring precise simulation results.
--inner_loop, --modeling_zone, and --outer_box specify the sizes of each respective zone in the mesh, defining the spatial extent of the mesh in these areas.
The element_size_inner, element_size_modeling, and element_size_outer options control the element size within each zone. Larger element sizes result in a coarser mesh and faster simulations, while smaller element sizes create a finer mesh with higher accuracy.
The --plane-z option allows specifying a reference plane above the modeling zone for advanced mesh configurations, with a default value of 0.0.
The element sizes within the modeling, and inner loop zones should be small enough to accurately capture details of the target cell for precise simulation results.
The mesh dimensions are unitless; no predefined units exist for any of the options. Users can interpret the dimensions according to their specific application needs.
The template_script specifies the base structure of the mesh, while the output_script name can be customized or left as the default (mesh.py).
The output mesh is saved to a file with a customizable name, defaulting to mesh.med.
generate-models: Creates physical models necessary for generating mesh and simulation data.
generate: Generates synthetic data based on the mesh and models created.
train: Trains AI models on the synthetic data generated using the mesh.
inference: Applies trained AI models to new data based on the synthetic structures.
