Wind Tunnel via Inductiva API

Wind tunnels are central for the development and testing of vehicle aerodynamics - from your everyday commuter car to the high-performance F1 car. With Inductiva API users can simplify their simulation workflow and build their custom virtual wind tunnels to accelerate the discovery of new designs.

In this repository, we assume users are familiar with one of the most popular open-source simulators for virtual wind tunnels, OpenFOAM. If not, we advise to get acquainted with it through the classical motorbike tutorial.

Virtual wind tunnel simulations

When performing wind tunnel tests - virtual or physical - only a few parameters are tweaked, for example, changing the airflow velocity for the same vehicle. Inductiva API provides a powerful mechanism to configure this simulation from this set of parameters, this workflow is designated as a simulation scenario.

Let's set an F1 car in a toy wind tunnel and run a simulation scenario:

import windtunnel

# Create wind tunnel object
wind_tunnel = windtunnel.WindTunnel(dimensions=(20,10,8))

# Set the object in the windtunnel
wind_tunnel.set_object(
    object_path="assets/f1_car.obj",
    rotate_z_degrees=0,
    normalize=False,
    center=True,
)

# Display the Windtunnel with the object
wind_tunnel.display()

# Submit a simulation task
task = wind_tunnel.simulate(wind_speed_ms=20,
                            num_iterations=50,
                            resolution=3)

# Wait for the task to finish
task.wait()

# Download all of the output files of the simulation
output_dir = task.download_outputs()

In this code snippet, we are using the WindTunnel to configure the simulation files for OpenFOAM, the F1 car and simulation parameters and simulating via Inductiva API. This scenario is a toy example that can be extended to more complex simulation scenarios.

For more detailed information, please visit the Wind Tunnel Documentation.

Installation

1. Clone the repository:

   git clone https://github.com/inductiva/wind-tunnel.git
   cd wind-tunnel

2. Install the package:

   pip install .

Scripts Overview

The repository includes several scripts to facilitate running and managing your wind tunnel simulations:

run.py

Quickly run standalone wind tunnel simulations using this script. It accepts the following arguments:

• --display: Opens a pyvista window with the windtunnel simulation
• --object_path: Path to the input object file
• --wind_speed_ms: Wind speed in meters per second (default: 10)
• --resolution: Resolution of the OpenFOAM simulation (default: 3, max: 5)
• --rotate_z_degrees: Rotation angle around the Z-axis in degrees (default: 0)
• --num_iterations: Number of iterations to run the simulation (default: 50)
• --machine_group_name: Machine group to run the simulation on (default: default queue)

Example usage:

python run.py --object_path assets/f1_car.obj --display

batch_run:

Submit multiple simulations with varying arguments. This script is useful for exploring different scenarios efficiently. Inside the script you can change the simulation arguments.

• --object_path: Path to the input object file
• --machine_group_name: Machine group to run the simulation on (default: default queue)

python batch_run.py --object_path assets/f1_car.obj 

view_outputs.py

Display the results of a completed simulation in a PyVista window and print the resulting force coefficients. This script waits for the task to finish and requires a task_id argument.

python view_outputs.py --task_id <task_id>

Running Simulations on Dedicated Hardware

To run simulations on dedicated hardware, you need to start a machine group.

First, list all available machine types with:

inductiva resources available

Starting a Machine Group

Note: Spot machines are preemptible but are approximately 5x cheaper. If a machine becomes unavailable during use, the simulation will automatically restart.

Instantiating a MachineGroup object with 5 preemptible machines of type c2-standard-32:

import inductiva

machine_group = inductiva.resources.MachineGroup(
    machine_type="c2-standard-30", num_machines=5, spot=True)
machine_group.start()

print(machine_group.name)

# Don't forget to terminate the machine group after using it:
machine_group.terminate()

Using the Machine Group

After starting the machine group, pass its name to the run.py or batch_run.py scripts.

python run.py --object_path assets/f1_car.obj --machine_group_name <machine_group_name>

Elastic Machine Groups

You can also create an ElasticMachineGroup, which scales machines up or down automatically based on the simulation queue.

For more details on running simulations on dedicated hardware, refer to the Inductiva Machine Group documentation

Disclaimer: For the sake of time the simulations performed here are low resolution and don't represent any practical real-world application.

To learn more about the inductiva package, check the Inductiva API documentation.