#*------------------------------------------------------------------------------*
#* JAX-FLUIDS - *
#* *
#* A fully-differentiable CFD solver for compressible two-phase flows. *
#* Copyright (C) 2022 Deniz A. Bezgin, Aaron B. Buhendwa, Nikolaus A. Adams *
#* *
#* This program is free software: you can redistribute it and/or modify *
#* it under the terms of the GNU General Public License as published by *
#* the Free Software Foundation, either version 3 of the License, or *
#* (at your option) any later version. *
#* *
#* This program is distributed in the hope that it will be useful, *
#* but WITHOUT ANY WARRANTY; without even the implied warranty of *
#* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the *
#* GNU General Public License for more details. *
#* *
#* You should have received a copy of the GNU General Public License *
#* along with this program. If not, see <https://www.gnu.org/licenses/>. *
#* *
#*------------------------------------------------------------------------------*
#* *
#* CONTACT *
#* *
#* deniz.bezgin@tum.de // aaron.buhendwa@tum.de // nikolaus.adams@tum.de *
#* *
#*------------------------------------------------------------------------------*
#* *
#* Munich, April 15th, 2022 *
#* *
#*------------------------------------------------------------------------------*
from typing import Callable, Dict
import jax
import jax.numpy as jnp
from jaxfluids.materials.material_manager import MaterialManager
from jaxfluids.solvers.riemann_solvers.riemann_solver import RiemannSolver
from jaxfluids.utilities import get_fluxes_xi
[docs]
class RusanovNN(RiemannSolver):
"""RusanovNN Riemann Solver which is optimied in the JAX-FLUIDS paper.
The dissipation is calculated by a neural network. See details in
Bezgin et al. 2022.
"""
def __init__(self, material_manager: MaterialManager, signal_speed: Callable) -> None:
super().__init__(material_manager, signal_speed)
[docs]
def solve_riemann_problem_xi(self, primes_L: jnp.ndarray, primes_R: jnp.ndarray,
cons_L: jnp.ndarray, cons_R: jnp.ndarray, axis: int,
ml_parameters_dict: Dict, ml_networks_dict: Dict, **kwargs) -> jnp.ndarray:
params = ml_parameters_dict["riemannsolver"]
net = ml_networks_dict["riemannsolver"]
# PHYSICAL FLUXES
fluxes_left = get_fluxes_xi(primes_L, cons_L, axis)
fluxes_right = get_fluxes_xi(primes_R, cons_R, axis)
# BUILD NEURAL NETWORK INPUTS
speed_of_sound_left = self.material_manager.get_speed_of_sound(p = primes_L[4], rho = primes_L[0])
speed_of_sound_right = self.material_manager.get_speed_of_sound(p = primes_R[4], rho = primes_R[0])
speed_of_sound = 0.5 * (speed_of_sound_left + speed_of_sound_right)
delta_vel = jnp.abs(primes_R[axis+1] - primes_L[axis+1])
mean_vel = 0.5 * (primes_L[axis+1] + primes_R[axis+1])
entropy_L = primes_L[4] / (primes_L[0])**self.material_manager.gamma
entropy_R = primes_R[4] / (primes_R[0])**self.material_manager.gamma
delta_s = jnp.abs(entropy_R - entropy_L)
# EVALUATE NEURAL NETWORK FOR RUSANOVNN DISSIPATION
vec = jnp.stack([delta_vel, mean_vel, speed_of_sound, delta_s])
dissipation = net.apply(params, vec)
# FINAL FLUX
fluxes_xi = 0.5 * ( (fluxes_left + fluxes_right) - dissipation * (cons_R - cons_L) )
return fluxes_xi