Tutorial: Using a File-Based Grid (Bent Channel)

Table of Contents

• 1. Case Overview: Laminar Flow in a Bend

In the previous tutorial, we ran a simulation where the grid was generated automatically. For complex, real-world geometries, you will typically use a dedicated meshing tool to create a grid and provide it to the solver as a file.

This tutorial will guide you through running the bent_channel example, which demonstrates how to use a pre-computed curvilinear grid file.

1. Case Overview: Laminar Flow in a Bend

This template simulates incompressible laminar flow through a 90-degree bent channel. The key feature is that the grid is read from a file, demonstrating the workflow for non-trivial geometries.

Physics:

• Flow Type: Laminar
• Reynolds Number (Re): 200
• Geometry: 90-Degree Bent Square Duct

Visualization of the bent channel grid file in ParaView.

1. Initializing the Study Directory

Just like before, we use the pic.flow init command to create our study directory.

1. Navigate to the root of your PICurv project directory.

2. Run the following command:

   bash ./bin/pic-flow init bent_channel --dest my_bent_channel_run

This copies the bent_channel template into a new my_bent_channel_run directory. If you look inside, you will see a new set of files:

my_bent_channel_run/
├── bent_channel.yml
├── Imp-MG-Standard.yml
├── Standard_Output.yml
├── standard_analysis.yml
├── bent_channel_coarse.picgrid  <-- The grid coordinate file
├── bent_channel_coarse.vts      <-- A VTK file for visualizing the grid
└── bent_channel_coarse.info     <-- Grid quality metrics (for reference)

2. Understanding the Configuration Files

Open the my_bent_channel_run/bent_channel.yml file and look at the grid section. It's different from the previous tutorial:

grid:
  mode: file
  source_file: bent_channel_coarse.picgrid

• **mode: file**: This tells pic.flow that the C-solver should not generate its own grid.
• **source_file: ...**: This provides the path to the coordinate file (.picgrid) that the solver should use.

Notice that the Imp-MG-Standard.yml and Standard_Output.yml files are identical to the ones from the flat_channel example. This highlights a core design principle of PICurv: physical cases, numerical solvers, and monitoring are modular and interchangeable.

3. Running the Simulation

The run command is almost identical to the previous tutorial. We just point the --case flag to our new bent_channel.yml file.

1. From the project root directory, execute the run command:

   bash ./bin/pic-flow run \ --case my_bent_channel_run/bent_channel.yml \ --solver my_bent_channel_run/Imp-MG-Standard.yml \ --monitor my_bent_channel_run/Standard_Output.yml \ --post my_bent_channel_run/standard_analysis.yml \ -n 4 --solve --post-process

2. The solver will start, read the grid file, and run the simulation.

5. Visualizing the Results

The output will again be in a timestamped directory inside runs/, with the final visualization files in the viz/ subdirectory.

1. Open the File Series in ParaView:
   • Go to File -> Open and select the Field..vts series from the viz/ directory.
   • Click Apply.
2. Visualize the Flow Path:
   • The best way to see the flow in a bend is with a Stream Tracer.
   • With Field..vts selected, click the Stream Tracer filter icon in the toolbar.
   • The "Seed" for the tracer is a point or line where streamlines will start. By default, it's a line. Drag the line so it is positioned at the inlet of your channel.
   • In the Properties panel, increase the number of "Resolution" points on the line to generate more streamlines.
   • Click Apply.
   • Color the resulting streamlines by the U_nodal vector field magnitude.

You will see the flow navigating the bend, with higher velocities on the inner wall and lower velocities on the outer wall, which is the expected physical result.

Example of streamlines in the bent channel, colored by velocity magnitude.

5. Next Steps

You have now successfully run a simulation using both a programmatic grid and a file-based grid. You understand the core user workflow of PICurv.Before moving on to create your own cases, it's essential to learn the fundamentals of visualizing and analyzing your results.

Please proceed to the final introductory tutorial: Tutorial: A Guide to Visualizing Your Results.