Skip to content

ansys/pysystem-coupling

Repository files navigation

PySystemCoupling

pyansys GH-CI codecov MIT black

Overview

PySystemCoupling provides Pythonic access to Ansys System Coupling. Although this Ansys product exposes its own Python-based scripting and command line interface, it is embedded and based on a specific version of Python. In contrast, PySystemCoupling enables seamless use of System Coupling within the Python ecosystem, providing additional capabilities, including:

  • Ability to launch System Coupling using a local Ansys installation
  • Access to APIs to set up and solve coupled analyses
  • Full access to the System Coupling data model via a convenient and Pythonic interface

Installation

Install PySystemCoupling with this command:

pip install ansys-systemcoupling-core

Alternatively, clone and install PySystemCoupling in development mode with this code:

git clone https://github.com/ansys/pysystem-coupling.git
cd pysystem-coupling
python -m pip install --upgrade pip
pip install -e .
pip install .[classesgen]
python scripts\generate_datamodel.py

Documentation and Issues

For more information, see the Documentation page.

Use the PySystemCoupling Issues page to post bug reports, questions and feature requests.

Usage

It is assumed that an Ansys installation is available and that this installation includes System Coupling and the participant products needed for the coupled analysis.

The System Coupling installation is found by examining the following environment variables in this order:

  • SYSC_ROOT
  • AWP_ROOT
  • AWP_ROOT242

If a variable is set but does not refer to a valid installation, PySystemCoupling fails at that point, rather than attempting to use the next variable.

In a standard user installation, the expectation is that only AWP_ROOT242 is set.

(It is also possible to provide a different version number as an argument to the launch() function. This will affect which AWP_ROOT<version> environment variable is examined.)

The System Coupling API is exposed to PySystemCoupling in two forms:

  • A documented interface based on concrete Python classes, following Pythonic conventions
  • A dynamic interface, undocumented in PySystemCoupling, that replicates the native System Coupling API

Both forms are strongly related to each other. A key difference in the Pythonic API is that naming is adjusted, in a generally predictable manner, to follow Python conventions. If you are already familiar with System Coupling, adjusting to this form, which is the recommended API, should be easy. However, if you are transitioning existing scripts, the native System Coupling API is made available as a convenience.

Note

While most commands should work as expected via the native System Coupling API, no guarantees can be given because of the nature of how it is exposed.

This example shows how to set up and solve an oscillating plate example in the Pythonic API. It uses Ansys Fluent as the CFD solver.

import ansys.systemcoupling.core as pysystemcoupling

syc = pysystemcoupling.launch()
setup = syc.setup
setup.add_participant(input_file="mapdl.scp")
setup.add_participant(input_file="fluent.scp")

## Create interfaces and data transfers by specifying participant regions

interface_name = "interface-1"
interface = setup.coupling_interface.create(interface_name)
interface.side["One"].coupling_participant = "MAPDL-1"
interface.side["One"].region_list = ["FSIN_1"]
interface.side["Two"].coupling_participant = "FLUENT-2"
interface.side["Two"].region_list = ["wall_deforming"]

# Use commands to add data transfers
force_transfer_name = setup.add_data_transfer(
    interface=interface_name,
    target_side="One",
    side_one_variable="FORC",
    side_two_variable="force",
)

disp_transfer_name = setup.add_data_transfer(
    interface=interface_name,
    target_side="Two",
    side_one_variable="INCD",
    side_two_variable="displacement",
)

# Change analysis duration and step size
setup.solution_control.time_step_size = "0.1 [s]"
setup.solution_control.end_time = "1.0 [s]"

# Set output control settings
setup.output_control.option = "StepInterval"
setup.output_control.output_frequency = 2

# Start streaming standard output from server
syc.start_output()

# Solve
solution = syc.solution
solution.solve()

The Pythonic API partitions commands via three high-level root attributes of the Session class: setup, solution, and case. The preceding example uses both the setup and solution attributes.

  • The setup attribute is the largest part of the API. It is where you find all commands related to populating the settings that define a coupled analysis. This attribute also provides direct access to the hierarchical data model.
  • The solution attribute is home to commands related to solving an analysis and examining the solution.
  • The case attribute, which is not used in the preceding example, provides all commands related to case file management and persistence.

While the preceding example uses the pysystemcoupling.launch() method to start the System Coupling server, alternatively, the server can be started in advance by calling command line arguments -m cosimgui --grpcport=<host:port> and pysystemcoupling.connect(host, port).

This next example shows how to set up the same analysis using the native System Coupling API. While the code here is less complete than the code shown previously, it should sufficiently illustrate the differences and connections between the two API forms.

import ansys.systemcoupling.core as pysystemcoupling

syc = pysystemcoupling.launch()
native_api = syc._native_api

native_api.AddParticipant(InputFile="mapdl.scp")
native_api.AddParticipant(InputFile="fluent.scp")

interface = native_api.CouplingInterface["interface-1"]
interface.Side["One"].CouplingParticipant = "MAPDL-1"
...

native_api.SolutionControl.TimeStepSize = "0.1 [s]"
...
syc.start_output()
native_api.Solve()

License

PySystemCoupling is licensed under the MIT license.

The ansys-systemcoupling-core package makes no commercial claim over Ansys whatsoever. It extends the functionality of Ansys System Coupling by adding a Python interface to the System Coupling service without changing the core behavior or license of the original software. Interactively controlling System Coupling via PySystemCoupling requires a local copy of System Coupling and licenses for all Ansys products involved in your coupled analysis.

To get a copy of Ansys, visit Ansys.