Compare full P3 with SP3 and P1#

Solve the same seven-fiber contact geometry on two skin layers with three angular approximations. Full P3 retains the three-dimensional spherical-harmonic moments through order three. The supplied values are illustrative, and agreement between the models does not establish accuracy against radiative transfer.

import os
import numpy as np
import MomentRT

ureg = MomentRT.ureg

show_figures = os.environ.get("MOMENTRT_SPHINX_GALLERY") != "1"

Define the common physical configuration#

The Henyey–Greenstein anisotropy belongs to the material. The three models share matched-index faces, six diffuse sources, mesh, and aperture. The finite mesh samples the layer interface at tetrahedron centroids, so the interface geometry and disk footprints also require mesh refinement.

dermis = MomentRT.Material(
    absorption=0.03 / ureg.millimeter,
    reduced_scattering=1.0 / ureg.millimeter,
    refractive_index=1.4,
    anisotropy=0.8,
    phase_function=MomentRT.phase_function.HenyeyGreenstein(),
)

epidermis = MomentRT.Material(
    absorption=0.06 / ureg.millimeter,
    reduced_scattering=1.6 / ureg.millimeter,
    refractive_index=1.4,
    anisotropy=0.7,
    phase_function=MomentRT.phase_function.HenyeyGreenstein(),
)

sample = MomentRT.StructuredMedium(
    origin=(-4, -4, 0) * ureg.millimeter,
    size=(8, 8, 4) * ureg.millimeter,
)

sample.add_background(material=dermis)

sample.add_layer(
    name="Epidermis",
    thickness=0.8 * ureg.millimeter,
    material=epidermis,
)

sources = []

for fiber_number in range(6):
    angle = fiber_number * np.pi / 3

    coordinates = np.array([np.cos(angle), np.sin(angle)]) * 1.8 * ureg.millimeter

    source = MomentRT.SurfaceSource.on_face(
        sample=sample,
        face="z_min",
        coordinates=coordinates,
        radius=0.8 * ureg.millimeter,
        wavelength=850 * ureg.nanometer,
        optical_power=(1 / 6) * ureg.milliwatt,
        name=f"Illumination fiber {fiber_number + 1}",
    )

    sources.append(source)

detector = MomentRT.SurfaceDetector.on_face(
    sample=sample,
    face="z_min",
    coordinates=np.array([0, 0]) * ureg.millimeter,
    radius=0.8 * ureg.millimeter,
    acceptance_angle=90 * ureg.degree,
    orientation="outward_normal",
    name="Collection fiber",
)

mesh = MomentRT.CartesianMesh.from_sample(sample=sample, shape=(9, 9, 7))

boundary = MomentRT.RobinBoundary(effective_reflectance=0)

boundaries = MomentRT.Boundaries.uniform(condition=boundary)

Solve and compare the detector signal#

P3 solves six even angular moments and reconstructs ten odd moments from their gradients. SP3 instead solves two composite fields. The contact detector currently uses a Lambertian angular-acceptance approximation for all three models, including P3. Each linear diffuse solve uses one source; the six fluences and detector powers are added incoherently. This does not model optical interference, fiber modes, or Fresnel entry losses.

solvers = {
    "P1": MomentRT.P1(boundaries=boundaries),
    "SP3": MomentRT.SP3(boundaries=boundaries),
    "P3": MomentRT.P3(boundaries=boundaries),
}

combined_fields = {}

for name, solver in solvers.items():
    results = []

    for fiber_number, source in enumerate(sources, start=1):
        simulation = MomentRT.Simulation(
            medium=sample,
            source=source,
            mesh=mesh,
            solver=solver,
            detectors={"contact": detector},
        )

        report = simulation.validate()

        if not report.valid:
            raise ValueError(report.summary())

        result = simulation.run()

        results.append(result)

        if name == "P1":
            print(
                f"Fiber {fiber_number} source area: sampled "
                f"{result.source_sampled_area.to('mm^2'):~P}, "
                f"geometric {result.source_geometric_area.to('mm^2'):~P}"
            )

    combined_values = sum(result.fluence.values for result in results)

    combined_fields[name] = MomentRT.ScalarField(
        mesh=results[0].fluence.mesh,
        values=combined_values,
        name=f"Combined {name} fluence rate",
        metadata={**results[0].fluence.metadata, "source_count": len(sources)},
    )

    collected_power = sum(result.detector_power["contact"] for result in results)

    residual_power = sum(result.energy_balance["residual_power"] for result in results)

    print(f"{name} collected power: {collected_power.to('mW'):~P}")

    print(f"{name} minimum fluence: {combined_values.min():~P}")

    print(f"{name} energy residual: {residual_power:~P}")

    print(
        f"{name} sampled detector area: {results[0].detectors['contact'].sampled_area.to('mm^2'):~P}"
    )

    print(
        f"{name} geometric detector area: {results[0].detectors['contact'].geometric_area.to('mm^2'):~P}"
    )
Fiber 1 source area: sampled 1.8333333333333333 mm², geometric 2.0106192982974678 mm²
Fiber 2 source area: sampled 2.1666666666666665 mm², geometric 2.0106192982974678 mm²
Fiber 3 source area: sampled 2.0 mm², geometric 2.0106192982974678 mm²
Fiber 4 source area: sampled 1.8333333333333333 mm², geometric 2.0106192982974678 mm²
Fiber 5 source area: sampled 2.1666666666666665 mm², geometric 2.0106192982974678 mm²
Fiber 6 source area: sampled 2.0 mm², geometric 2.0106192982974678 mm²
P1 collected power: 0.06642209718791982 mW
P1 minimum fluence: -1.3068580830308083 W/m²
P1 energy residual: -1.3552527156068805e-18 W
P1 sampled detector area: 1.6666666666666665 mm²
P1 geometric detector area: 2.0106192982974678 mm²
SP3 collected power: 0.06471650491678295 mW
SP3 minimum fluence: -1.0599514053421157 W/m²
SP3 energy residual: -8.944667923005412e-19 W
SP3 sampled detector area: 1.6666666666666665 mm²
SP3 geometric detector area: 2.0106192982974678 mm²
P3 collected power: 0.05794231277494402 mW
P3 minimum fluence: -1.6575186027265618 W/m²
P3 energy residual: 1.2739375526704677×10⁻¹⁸ W
P3 sampled detector area: 1.6666666666666665 mm²
P3 geometric detector area: 2.0106192982974678 mm²

Explore the full P3 fluence#

Linear scale retains any negative values from the angular or spatial approximation. Refine the mesh and compare against an independent transport reference before interpreting the result for a particular sensor. The gallery preview is captured from this same combined P3 volume.

overlays = [*sources, detector]

fluence_figure = combined_fields["P3"].plot(
    view="volume",
    scale="linear",
    overlays=overlays,
    show=show_figures,
)

Inspect the modeled geometry#

This view shows the configured geometry and instrument footprints.

gallery_system_figure = sample.plot(overlays=overlays, show=False)

Total running time of the script: (0 minutes 2.205 seconds)

Gallery generated by Sphinx-Gallery