Join our global community on Discord and let's shape the next wave of simulation and scientific computing Let's shape the future of simulation together

Join us on Discord

Generalizing the Use Case

Inductiva enables the parallel execution of numerous simulations, making it ideal for conducting parameter studies at scale.

For example, suppose you want to analyze how varying a specific input parameter affects the simulation results. To demonstrate, consider studying the impact of changes in the mean wind velocity at the reference height, denoted as URef. This parameter is defined in the Meteorological Boundary Conditions section of the 90m_12mps_twr.inp input file.

In the current configuration, URef is set to 12 m/s.

--------Meteorological Boundary Conditions-------------------
"IECKAI"      TurbModel       - Turbulence model
"1-ed3"       IECstandard     - Number of IEC 61400-x standard
"B"           IECturbc        - IEC turbulence characteristic
"NTM"         IEC_WindType    - IEC turbulence type 
"default"     ETMc            - IEC Extreme Turbulence Model "c" parameter [m/s]
"PL"          WindProfileType - Velocity profile type 
         90   RefHt           - Height of the reference velocity (URef) [m]
         12   URef            - Mean (total) velocity at the reference height [m/s]
"default"     ZJetMax         - Jet height [m]
"default"     PLExp           - Power law exponent [-]
"default"     Z0              - Surface roughness length [m]

We will use the Inductiva templating system to generalize the base simulation by randomly sampling values for URef, RandSeed1, and RandSeed2. In the next section, we will use this approach to generate a diverse dataset of simulations.

Parametrize the `90m_12mps_twr.inp` file

Inductiva allows you to convert fixed parameters in your simulation configuration files into variables that you can set programmatically using Python scripting. This means that we will be able to change the URef, RandSeed1, and RandSeed2 defined in the 90m_12mps_twr.fst input file from a Python script before starting the simulation.

To do this, you will need to edit your 90m_12mps_twr.inp file from this:

---------Runtime Options-----------------------------------
...
      13428  RandSeed1        - First random seed
   "RanLux"  RandSeed2        - Second random seed
...
--------Meteorological Boundary Conditions-------------------
...
         12  URef             - Mean (total) velocity at the reference height [m/s]
...

To this:

---------Runtime Options-----------------------------------
...
   {{ seed_1 }}  RandSeed1    - First random seed
   {{ seed_2 }}  RandSeed2    - Second random seed
...
--------Meteorological Boundary Conditions-------------------
...
   {{ URef }}    URef         - Mean (total) velocity at the reference height [m/s]
...

After this small edit, you will need to save your input file as 90m_12mps_twr.inp.jinja (note the .jinja extension).

This will tell Inductiva's templating engine that Python variables called URef, seed_1 and seed_2 should be used to set the correct scalar value in 90m_12mps_twr.fst.

Code Overview

The script below shows how we can now set the value of the URef, RandSeed1, and RandSeed2 parameters from Python and run a variation of the original simulation:

import inductiva

# Allocate cloud machine
cloud_machine = inductiva.resources.MachineGroup(
   provider="GCP",
   machine_type="n2-highcpu-2",
   spot=True
)

seed_1 = 3
seed_2 = 5
URef = 15

print(f"Preparing files for TurbSim with "\
      f"seed1 = {seed_1}, seed_2 = {seed_2}, "\
      f"URef = {URef}")
target_dir = f"variations/s1_{seed_1}/s2_{seed_2}/URef_{URef}"

# This is where we make the substitution and set
# the value of the parameters in the modified config file
# turbsim 90m_12mps_twr.inp.jinja
inductiva.TemplateManager.render_dir(
   source_dir="input_files",
   target_dir=target_dir,
   overwrite=True,
   seed_1=seed_1,
   seed_2=seed_2,
   URef=URef)

# Initialize OpenFAST simulator
turbsim = inductiva.simulators.OpenFAST()

task = openfast.run(
   input_dir=target_dir,
   commands=["turbsim 90m_12mps_twr.inp"],
   on=cloud_machine)

task.wait()
cloud_machine.terminate()

First, we set the values for seed_1, seed_2 and URef. Then, we generate the input file using the render_dir method, which replaces the placeholder variables in our 90m_12mps_twr.fst.jinja template with the corresponding values.

The generated input files are then stored in the folder target_dir (variations/s1_{seed_1}/s2_{seed_2}/URef_{URef}).

For more details on managing template files, check out the TemplateManager documentation.

The good thing about the Inductiva API is that it’s a Python API, so you can simply use a for loop with random sampling for URef, seed_1 and seed_2 to run multiple simulations in parallel and generate a dataset. That’s exactly what we’ll do in the next section.

Run a Single Simulation

Generate the Dataset

Generalizing the Use Case

Parametrize the 90m_12mps_twr.inp file

Code Overview

Parametrize the `90m_12mps_twr.inp` file