This tutorial is presented as Python code running inside a Jupyter Notebook, the recommended way to use Salvus. To run it yourself you can copy/type each individual cell or directly download the full notebook, including all required files.

Reading External Meshes


Sometimes it is useful to interact with third-party meshing software. This may be case, for example, when using meshes based on CAD models. Salvus supports the use of such meshes, and has automated readers for a collection of open-source mesh formats. In this tutorial we'll give an overview of the currently supported formats along with an example of their use. The first thing to do of course is to import the required Python packages.

%matplotlib inline

from pathlib import Path

import h5py
import matplotlib.pyplot as plt
import numpy as np
import os
import salvus.mesh.unstructured_mesh as um
from salvus.flow import api
from salvus.flow import simple_config as sc
from typing import List

SALVUS_FLOW_SITE_NAME = os.environ.get("SITE_NAME", "local")

Also, to make our lives a bit easier, we'll write a few short functions that will help us quickly generate Salvus inputs as we proceed

def get_basic_source(
    *, frequency: float, physics: str, location: List
) -> sc.source:
    """Gets a simple physics- and dimension-dependent source.

        frequency: Center frequency of the (Ricker) wavelet.
        physics: Physics of the source.
        location: Location of the source.

        SalvusFlow source object appropriate for the specified physics.

    l = location
    src = sc.source.cartesian
    s = sc.stf.Ricker(center_frequency=frequency)
    if physics == "acoustic":
        return src.ScalarPoint3D(
            x=l[0], y=l[1], z=l[2], f=1, source_time_function=s

    return src.VectorPoint3D(
        x=l[0], y=l[1], z=l[2], fx=1.0, fy=1.0, fz=0.0, source_time_function=s

Exodus is one of the more common third-party mesh formats, and SalvusMesh can read Exodus files natively. In the following example we'll read some Exodus meshes into Salvus and run some simulations. We'll focus on two use cases: a purely elastic simulation, and a coupled fluid-solid simulation.


# Read mesh from Exodus file.
mesh = um.UnstructuredMesh.from_exodus("./data/glass.e")

# Find the surface of mesh.

# Mark simulation as elastic.
mesh.attach_field("fluid", np.zeros(mesh.nelem))

# Attach parameters.
pars = {"VP": 5800, "VS": 4000, "RHO": 2600}
template = np.ones_like(mesh.get_element_nodes()[:, :, 0])
for key, value in pars.items():
    mesh.attach_field(key, template * value)

# Visualize.
<salvus.mesh.unstructured_mesh.UnstructuredMesh at 0x7fbd4a0eee10>
# Set up simulation.
w = sc.simulation.Waveform(
        frequency=100.0, physics="elastic", location=[0, 0, 70]

# Generate a movie.
w.output.volume_data.format = "hdf5"
w.output.volume_data.filename = "movie.h5"
w.output.volume_data.fields = ["displacement"]
w.output.volume_data.sampling_interval_in_time_steps = 100

# Validate simulation parameters.
Job `job_2101132307202971_c02a6a9d78` running on `local` with 2 rank(s).
Site information:
  * Salvus version: 0.11.24
  * Floating point size: 32
* Downloaded 42.1 MB of results to `output`.
* Total run time: 8.46 seconds.
* Pure simulation time: 8.06 seconds.
<salvus.flow.sites.salvus_job.SalvusJob at 0x7fbd41ae5ad0>

Snapshot of the magnitude of dynamic elastic displacement, visualized using Paraview and the output .xdmf file.

# Read mesh from Exodus file.
mesh_solid = um.UnstructuredMesh.from_exodus("./data/solid.e")
mesh_fluid = um.UnstructuredMesh.from_exodus("./data/fluid.e")

# Mark fluid and solid elements.
mesh_solid.attach_field("fluid", np.zeros(mesh_solid.nelem))
mesh_fluid.attach_field("fluid", np.ones(mesh_fluid.nelem))

# Attach parameters (elastic).
pars = {"VP": 5800, "VS": 4000, "RHO": 2600}
template = np.ones_like(mesh_solid.get_element_nodes()[:, :, 0])
for key, value in pars.items():
    mesh_solid.attach_field(key, template * value)

# Attach parameters (acoustic).
pars = {"VP": 1500, "VS": 0, "RHO": 1000}
template = np.ones_like(mesh_fluid.get_element_nodes()[:, :, 0])
for key, value in pars.items():
    mesh_fluid.attach_field(key, template * value)

# Combine both element blocks.
mesh = mesh_solid + mesh_fluid

# Find the surface of mesh.

# Visualize.