space-rha.comShort courseChapter 2 · Shielding and transport

Chapter 2 · Lesson 2 of 6 · about 10 minutes

Shielding has directions

A part near an opening sees much less material in one direction than in another. Moving the part, or adding a real cover, changes those paths. A single wall thickness cannot describe the surroundings of every part in the box.

Shielding has directions

Figure 2.2. Dose against aluminum thickness at 800 km and 98.6°, by particle (chapter 1, lesson 5). A real part sees a different thickness in every direction; sector analysis weights this curve by the solid angle of each direction.
Polar LEO 800 km: one year of dose behind aluminum, by particle 0.1 1 10 1 10 100 1000 104 105 aluminum shielding, g/cm² (1 g/cm² = 3.7 mm = 146 mil) rad(Si) per year, center of a solid sphere 100 mil trapped electrons + bremsstrahlung trapped protons one August 1972 solar event bremsstrahlung alone

Placement

Where a part sits in the box changes the material around it. A part at the center of the enclosure and a part next to an opening see different path lengths in every direction. Moving the hardware is a trade worth considering alongside adding material.

Coverage

A thick wall on one side does not close an opening on another. Which directions matter depends on the incident field and on the response you are calculating.

What a path histogram shows

A histogram of path lengths through the mass model measures the material around the target in each direction. It is a geometric quantity. It says nothing yet about dose or shielding effectiveness; those need a response or a transport calculation on top of it.

Mass thickness and material paths

A ray traced through the mass model records the physical length through each solid and the order of the materials it crosses. Multiplying each length by the material density gives the areal mass along the ray. Straight rays are good at finding coverage and openings. They know nothing about scattering or secondary production, so those need a response library or a transport model.

Areal mass. Along a path, x = ∫ρ ds. For a uniform layer this is x = ρℓ, with length in cm and density in g/cm³. Two shields with equal areal mass are equal in mass, and that is all the number says. Whether they are equivalent for a given particle and effect is a separate question.

Slab approximation. A ray crossing an ideal plane layer of thickness t at angle θ to its normal has path length t/cos θ. A real box adds side walls, seams and openings that the slab formula does not see. Comparing the slab estimate with the traced paths through the declared enclosure shows how much the idealization misses.

Material sequence. Keep each material as its own column, in the order the ray crosses them, when the downstream method can use that information. Collapsing everything to an aluminum equivalent throws away what the method would need to treat electron, photon and nuclear interactions in the actual materials.

Angular sample. A simple ray set samples the full 4π sphere with roughly equal weights. That describes the geometry, and only the geometry. A mission source may arrive anisotropically, or need weights that depend on energy, and the ray set then has to be weighted to match it.

3D scene

Shielding has directions

Rays from the target through the enclosure: wall, opening, cover and board. Change the wall and the opening and watch the sampled mass thickness change by direction.

  1. Rays traced from the target through the mass model record the length through each solid and the order of the materials crossed. The histogram of those path lengths measures the material around the target in each direction. It is a geometric quantity; it says nothing yet about dose or shielding effectiveness.

  2. With the opening widened, the rays through it meet little or no material. A thick wall on one side does not close an opening on another, and which directions matter depends on the incident field and on the response being calculated. Moving the part away from the opening is a trade worth considering alongside adding material.

  3. Thickening the wall from 2.5 mm to 8 mm raises the areal mass along every ray that crosses it and leaves the rays through the opening unchanged. Along a path x = ∫ρ ds; for a uniform layer x = ρℓ, and a ray at angle θ to the normal of an ideal plane layer crosses t/cos θ of it. The traced paths show how much the slab formula misses in a real box with side walls, seams and openings.

  4. Tantalum at the same thickness carries far more areal mass on every ray. Two shields with equal areal mass are equal in mass, and that is all the number says; whether they are equivalent for a given particle and effect is a separate question, which is why the material sequence along each ray should be kept when the downstream method can use it.

The original space-rha.com environment and mission renderer (NASA imagery and spacecraft models, credits in the view), with the course scenes drawn by the same renderer. Trails, clocks and counts are illustrative; the scene computes no dose or rate.

Exercise

Turn a physical mass model into directional shielding inputs for a response calculation, keeping clear that ray tracing is geometry and not particle transport.

When you connect the geometry to a response model, keep the direction, material and path conventions unchanged, because the response library was built for a particular convention.

Going deeper

Angular response and geometry error

A sector calculation folds the path length in each direction with a dose-depth response, under stated approximations. If you average the geometry first and then apply the response once, the answer can change. Whether that shortcut is adequate depends on how thin the thinnest paths are, how anisotropic the source is, how much lateral scattering there is and whether the materials are mixed.

Response before averaging. For a single-material response that matches the geometry convention, the sector sum has the form D ≈ ΣᵢwᵢDᵢ, with one dose-depth value per direction. Reading the response once at the mean thickness generally gives a different number. The weights have to represent the source and the angular convention you intend.

Sampling. A sampled opening fraction of zero means that no ray found an opening. It does not prove the box is closed. Increase the angular resolution, aim rays at the directions you are suspicious of, and check the small features. A ray set that has converged for a uniform source may still be badly sampled for a sharply directional one.

Transport beyond paths. Scattering, backscatter, secondaries that travel sideways and material interfaces all couple one direction to another, and a straight ray cannot represent that. Check the ray-response approximation wherever those contributions could change the engineering conclusion.

Geometry evidence. Check the model's scale, topology, overlaps, shells, densities and total mass against controlled CAD and the hardware itself.

Takeaway

Look around the target before adding shielding mass.

Check yourself

When is a sector analysis with slab transport good enough, and when is it not?

It is generally adequate for total dose from a smooth spectrum through aluminum-like materials, where the dose depends mostly on the areal density along each ray. It is generally not adequate where secondary particles, high-atomic-number materials or the energy dependence of the response matter, and for heavy-ion LET spectra, where the thin directions decide the answer and the approximations break down.

A box has one thin wall facing deep space and five thick ones. Which direction sets the dose, and what does that say about adding shielding?

The thin wall dominates, because the dose through a direction falls steeply with thickness; adding mass to the thick walls changes little, while adding it to the thin wall changes a lot. Shielding is bought where the rays are thin, which a sector analysis shows directly.

A ray trace reports an opening fraction of zero and the designer concludes that the box is closed. What does the zero show?

Only that no ray in that set found an opening; it does not prove the box is closed. The remedy is to increase the angular resolution, aim rays at the directions you are suspicious of and check the small features, because a ray set that has converged for a uniform source may still be badly sampled for a sharply directional one. The same model should also be checked for scale, topology, overlaps, shells, densities and total mass against the controlled CAD and the hardware itself.

To save time, the path lengths are averaged over all directions and the dose-depth curve is read once at the mean thickness. Why does that generally give a different number from the sector sum, and what decides whether the shortcut is adequate?

The sector sum applies the response in each direction first, D ≈ ΣᵢwᵢDᵢ, and then weights the results. Because the dose through a direction falls steeply with thickness, the thin directions contribute far more than their share of the solid angle, and a mean thickness does not represent them. Whether the shortcut is adequate depends on how thin the thinnest paths are, how anisotropic the source is, how much lateral scattering there is and whether the materials are mixed.

A ray-trace tool offers to collapse every path to an aluminum-equivalent thickness. When is that a loss, and what should be kept instead?

Collapsing to an aluminum equivalent throws away the material sequence that a downstream method would need to treat electron, photon and nuclear interactions in the actual materials; two shields of equal areal mass are equal in mass, and that is all the number says. Keep each material as its own column, in the order the ray crosses them, whenever the downstream method can use that information. For a total-dose sector sum through aluminum-like materials the collapse is generally adequate, and the loss matters where high-atomic-number materials, secondaries or the energy dependence of the response enter (lesson 2.3).

Further reading