Skip to main content

MPI Interface for m-AIA CFD Solver

This module provides the MPI MPMD (Multiple Program Multiple Data) interface for communication between Python RL controllers and the m-AIA CFD solver.

MaiaInterface Objects​

class MaiaInterface()

MPI MPMD interface for communication with the m-AIA CFD solver.

This class provides an interface to m-AIA via the MPI multiple program multiple data (MPMD) execution model. The m-AIA binary and Python controller must be launched together, sharing MPI_COMM_WORLD.

Example launch command: mpirun -np <n_controller_ranks> <controller> : -np <n_maia_ranks> <maia>

Attributes:

  • nDim - Number of spatial dimensions (2 or 3).
  • worldComm - MPI world communicator.
  • appComm - Application-specific communicator.
  • appRank - Rank within the application communicator.
  • remoteRoot - Root rank of the remote (m-AIA) application.

__init__​

def __init__(nDim: int)

Initialize the MaiaInterface.

Arguments:

  • nDim - Number of spatial dimensions (2 or 3).

init_comm​

def init_comm(comm_world: MPI.Comm) -> None

Initialize MPI communication with m-AIA.

Sets up the communicators and determines the root ranks for both the Python controller and the m-AIA solver.

Arguments:

  • comm_world - MPI communicator, typically MPI.COMM_WORLD.

continueRun​

def continueRun() -> None

Signal m-AIA to continue the simulation run.

This sends a continue command to allow the CFD solver to proceed after data exchange.

runTimeSteps​

def runTimeSteps(time_steps: int = 1) -> None

Advance the m-AIA simulation by a specified number of time steps.

Arguments:

  • time_steps - Number of time steps to advance. Use 0 to signal finish.

finishRun​

def finishRun() -> None

Signal m-AIA to finish the simulation run.

This sends a time step count of 0, indicating the run should terminate.

setControlProperties​

def setControlProperties(control_actions: Union[List, np.ndarray]) -> None

Set control properties for boundary condition actuation.

Arguments:

  • control_actions - Control action values. For jets in 2D with n jets, format is [u0, v0, u1, v1, ..., un, vn].

getProbeData​

def getProbeData(probe_point_coords: Union[List, np.ndarray]) -> np.ndarray

Get flow field data at specified probe locations.

Arguments:

  • probe_point_coords - Probe coordinates. For 2D with 3 probes, format is [x0, y0, x1, y1, x2, y2].

Returns:

Array of probe data. For 2D: [u, v, rho, p] per probe. For 3D: [u, v, w, rho, p] per probe.

getForce​

def getForce(bc_segment_id: int) -> np.ndarray

Get the force acting on a boundary segment.

Arguments:

  • bc_segment_id - Index of the boundary condition segment.

Returns:

Force vector array of length nDim.

reinit​

def reinit() -> None

Reinitialize the m-AIA simulation.

Triggers the m-AIA solver to recall its initialization routine, resetting the simulation to its initial state.