8 SpaceX Thermal Engineer Interview Questions & Answers

I would model the dominant assumed heat path from the dish electronics to an isothermal mount or support boundary as three series resistances. Starting at T_hot, heat Q̇ passes through conduction resistance R_cond in the dish structure, contact resistance R_contact at the mounting interface, and geometry dependent spreading resistance R_spread into the mount or support, ending at T_boundary. For steady state heat flow, R_total = R_cond + R_contact + R_spread and Q̇ = (T_hot − T_boundary) / R_total. Because the same heat rate passes through each series resistance, ΔT_i = Q̇R_i, so the largest resistance produces the largest temperature drop.

A useful first model is a single dominant heat path from the electronics heat source to an assumed isothermal mount or support sink. This is a conceptual thermal resistance circuit, not a claim about the exact internal construction of a Starlink dish.

1. Should I assume steady state operation and one dominant conduction path?
2. Should the mount or support be treated as an isothermal temperature boundary?
3. Do you want only the steady state resistance circuit, or should I also discuss transient checks using Biot and Fourier numbers?

Start with the heat source at T_hot and define positive heat flow Q̇ toward the support boundary at T_boundary.

Model the dominant path with three series resistances.

R_cond represents conduction through the dish structure. For a simple slab approximation, R_cond = L/(kA), where L is the conduction path length in m, k is thermal conductivity in W/(m K), and A is cross sectional area in m². Conductivity k is a material property, while thermal resistance depends on both the material and geometry.

R_contact represents the mounting interface. If h_c is the thermal contact conductance in W/(m² K), then R_contact = 1/(h_cA). Contact conductance describes how readily heat crosses the interface, while contact resistance is its resistance representation for the selected area.

R_spread represents heat spreading into the mount or support. It depends on geometry and material properties, so I would not assign a numerical value without the actual design geometry.

For the series path: R_total = R_cond + R_contact + R_spread

The steady state heat rate is: Q̇ = (T_hot − T_boundary)/R_total

The temperature drop across each element is: ΔT_i = Q̇R_i

Therefore, the largest series resistance produces the largest temperature drop.

The model assumes steady state behavior, one dominant conduction path, and properties evaluated at a relevant temperature. I would check actual geometry, conductivity, contact conductance, the support boundary condition, temperature dependence, anisotropy, and possible parallel heat paths. If transient behavior matters, I would use the Biot number to judge whether a lumped temperature approximation is reasonable and the Fourier number to characterize transient diffusion through the solid. Radiation or convection would require additional paths if they are significant.

Common mistakes are replacing the entire assembly with one resistance, omitting contact or spreading resistance, inventing a numerical spreading resistance without geometry, confusing thermal conductivity with thermal conductance or resistance, drawing heat flow in the wrong direction, or using a boundary temperature without stating the boundary assumption. Another mistake is applying a steady state or lumped approximation to an important transient case without checking whether that approximation is justified.

Draw the heat source and thermal boundary first, then place each resistance along the dominant heat path. Label R_cond, R_contact, and R_spread, state that the path is an assumed conceptual model, and finish with R_total, Q̇, and ΔT_i = Q̇R_i. This shows both the circuit and the physical reasoning without inventing internal construction details.

Reynolds number is a dimensionless measure of the relative importance of inertial effects compared with viscous effects in a fluid flow. It is commonly written as Re = ρVL/μ = VL/ν. Low Reynolds number means viscous effects are relatively more important and laminar behavior is more likely. High Reynolds number means inertial effects are relatively more important and turbulent behavior is more likely. The characteristic length depends on the geometry, such as pipe diameter for flow through a circular pipe.

Reynolds number compares the relative importance of fluid inertia with viscous resistance. It is dimensionless, so it is useful for comparing flows with different sizes, speeds, fluids, and operating conditions.

For characteristic velocity V and characteristic length L,

Re = ρVL/μ = VL/ν

where ρ is fluid density in kg/m³, V is characteristic velocity in m/s, L is characteristic length in m, μ is dynamic viscosity in Pa·s, and ν is kinematic viscosity in m²/s.

For internal flow through a circular pipe, the characteristic length is the pipe diameter D, so Re = ρVD/μ.

1. Are we considering internal pipe flow or an external flow geometry?
2. What characteristic velocity and length should I use for this geometry?
3. Should the fluid properties be evaluated at a specified operating temperature?

Start with the physical meaning: Reynolds number tells me whether inertial effects or viscous effects are more influential in the flow.

At low Re, viscous effects dominate. Disturbances are more strongly damped, so laminar behavior is more likely. For fully developed Newtonian flow in a circular pipe, the laminar velocity profile is parabolic.

At high Re, inertial effects are relatively stronger. Disturbances can persist more readily, so turbulent behavior is more likely.

I would not treat Reynolds number as a universal transition switch. The actual transition depends on geometry, disturbances, surface condition, inlet conditions, and other boundary conditions. That is why two cases with the same geometry can show different flow behavior when their Reynolds numbers differ.

For circular internal pipe flow, I would use the pipe diameter as the characteristic length. If velocity increases while density, viscosity, and diameter remain unchanged, Reynolds number increases. If viscosity increases while the other terms remain unchanged, Reynolds number decreases.

The engineering use is to characterize the likely flow regime and choose models or correlations that apply to that regime and geometry. Before using a friction factor relation, pressure drop model, or convection correlation, I would check that the Reynolds number and geometry fall within the documented applicability range and consider entrance effects where relevant.

So the key conclusion is that Reynolds number measures the relative importance of inertial effects to viscous effects in fluid flow.

Common mistakes include saying Reynolds number directly measures turbulence, using the wrong characteristic length, ignoring how operating conditions affect fluid properties, assuming one transition value applies to every geometry, or using a heat transfer or pressure drop correlation outside its documented Reynolds number range.

Lead with the physical meaning before giving the equation: Reynolds number compares inertial effects with viscous effects. Then define the symbols and give one simple geometry example, such as using pipe diameter for circular internal flow. Mention that transition depends on geometry and disturbances rather than presenting Reynolds number as an absolute turbulence switch.

The Stefan Boltzmann law states that the total emissive power of a blackbody is Eb = σT^4, where T is absolute temperature in kelvin and σ is the Stefan Boltzmann constant, approximately 5.67 × 10^−8 W m^−2 K^−4. A real gray surface emits εσT^4 per unit area, where ε is emissivity. For a diffuse gray surface completely surrounded by a large black enclosure, with α = ε and Fs→sur = 1, the net outward radiation rate is q̇rad = εσA(Ts^4 − Tsur^4). The fourth power dependence makes thermal radiation very sensitive to absolute temperature.

A simple way to explain the Stefan Boltzmann law is to start with an ideal blackbody, then extend it to a real surface and finally to net radiation exchange with the surroundings.

1. Should I explain only surface emission, or also net radiation exchange with the surroundings?
2. May I assume a diffuse gray surface inside a large black enclosure?
3. Should I discuss spectral effects, or use total hemispherical properties?

For an ideal blackbody, the Stefan Boltzmann law gives the total emissive power as Eb = σT^4. Here Eb is in W/m², T is the absolute temperature in kelvin, and σ ≈ 5.67 × 10^−8 W m^−2 K^−4.

For a real gray surface, emission is reduced by emissivity ε, so a surface of area A at temperature Ts emits q̇emit = εσATs^4. Emissivity is dimensionless and lies between 0 and 1.

For the diagrammed case, the surface is diffuse and gray and is completely surrounded by a large black enclosure at temperature Tsur. The geometric assumption is Fs→sur = 1. Under the gray surface assumption, absorptivity equals emissivity for the corresponding thermal radiation conditions, so α = ε. The absorbed radiation from the enclosure is therefore q̇abs = εσATsur^4.

The net outward radiation rate is the emitted rate minus the absorbed rate:

q̇rad = εσA(Ts^4 − Tsur^4).

This sign convention is useful. If Ts > Tsur, the surface loses heat by radiation. If Ts = Tsur, the net exchange is zero even though both bodies continue to emit and absorb radiation. If Ts < Tsur, q̇rad is negative, so the surface has a net radiative heat gain.

The most important physical point is the fourth power dependence. Doubling the absolute temperature increases blackbody emissive power by a factor of 16. Temperature must therefore be used in kelvin.

Emissivity, absorptivity, and reflectivity should not be treated as interchangeable constants in general. For an opaque surface, α + ρ = 1 at a given wavelength and direction. For nongray surfaces, wavelength dependent properties may be needed.

For more general geometries, I would use view factors and check enclosure relationships such as summation and reciprocity. Radiation is also distinct from convection because it does not require a material medium and can occur across vacuum.

Common mistakes include using degrees Celsius instead of kelvin, forgetting the fourth power on temperature, applying εσAT^4 as a net heat transfer equation without accounting for radiation received from the surroundings, confusing emissivity with absorptivity or reflectivity when their relationship is not justified, ignoring view factors when surfaces do not completely see each other, and applying one constant emissivity when spectral variation is important.

State the blackbody law first, then extend it to a real gray surface and finally to net exchange with the surroundings. Define every symbol, state that temperature must be absolute, explain the sign of Ts^4 − Tsur^4, and identify the assumptions Fs→sur = 1 and α = ε that make the simplified enclosure equation valid.

A spacecraft radiator operating in vacuum is a strong example. Electronics generate waste heat, Q̇in, which is conducted through the spacecraft thermal path to an external radiator. Because vacuum provides essentially no convective heat transfer, the radiator rejects that energy primarily by thermal radiation to its surroundings. For a simplified radiator with a clear view of deep space, Q̇rad = εσAF(Tr⁴ − Tspace⁴), where ε is infrared emissivity, A is radiator area, F is the view factor to deep space, and both temperatures are absolute temperatures in Kelvin.

1. Should I assume the radiator has a clear view of deep space?
2. Should solar and planetary radiation be neglected for this example?
3. Can I treat the radiator as a diffuse gray surface?

A good application is an external spacecraft radiator operating in vacuum. Electronics inside the spacecraft generate waste heat, Q̇in. That heat can still move through solid structures by conduction, so it is conducted to the radiator surface. At the external radiator, convection is essentially unavailable because there is no surrounding fluid. The radiator therefore rejects heat primarily through thermal radiation.

For the simplified case shown, define the radiator surface temperature as Tr and the effective deep space temperature as Tspace, both in Kelvin. Let ε be the radiator infrared emissivity, A the radiating area, F the view factor to deep space, and σ the Stefan Boltzmann constant, approximately 5.67 × 10⁻⁸ W m⁻² K⁻⁴. The net radiative heat rejection can then be represented as

Q̇rad = εσAF(Tr⁴ − Tspace⁴).

This relation shows why absolute temperature matters. Radiative emission varies with the fourth power of temperature, so a hotter radiator can reject substantially more heat for the same area and emissivity. A high infrared emissivity also increases thermal emission and therefore improves heat rejection for the same geometry and temperature.

The view factor F represents geometric visibility between the radiator and deep space. The simplified expression is appropriate when the radiator has a clear view of that radiative sink and other incident radiation is neglected. If the radiator also sees a warm planet or another spacecraft surface, those surfaces must be included separately with their own temperatures and view factors.

Surface properties should also be interpreted carefully. Emissivity describes thermal emission, absorptivity describes absorption of incident radiation, and reflectivity describes the reflected fraction. They should not automatically be treated as interchangeable constants across different wavelength ranges. Solar radiation and infrared thermal radiation can therefore require different surface property values.

As a check, I would verify that temperatures are in Kelvin, confirm the radiator energy balance between incoming conducted heat and outgoing net radiation, and check that the relevant view factors represent the actual geometry. The key conclusion is that conduction transports the electronics heat to the radiator, while radiation is the external heat rejection mechanism to space.

Common mistakes include saying that conduction cannot occur in vacuum, even though heat can still conduct through solid spacecraft structures, using Celsius instead of Kelvin in the Stefan Boltzmann relation, ignoring the radiator view factor, assuming every radiator sees only deep space, and treating emissivity, absorptivity, and reflectivity as interchangeable surface properties.

Start with the physical environment. State that the spacecraft is in vacuum, so external convection is essentially absent. Then trace the same heat path shown in the diagram: electronics generate Q̇in, conduction moves that heat to the radiator, and the radiator rejects Q̇rad by thermal radiation to deep space. Finish by mentioning high infrared emissivity, absolute temperature, radiator area, and view factor.

Window condensation occurs when humid indoor air contacts glass whose interior surface temperature is below the indoor air dew point. Heat flows from the warmer room toward the colder exterior, cooling the glass and the adjacent air. When that local air reaches saturation, water vapor changes phase into liquid on the glass and releases latent heat. The droplets can grow, merge, and run downward. If the surface remains above the dew point, condensation does not occur.

Window condensation is a surface phase change problem. The working fluid is water contained as vapor in indoor air. In the winter condition shown, the indoor air is warmer and more humid than the outdoor air, while heat flows through the window toward the colder exterior. This lowers the temperature of the interior glass surface. Condensation begins when the local surface temperature, T surface, falls below the indoor air dew point temperature, T dew point. Air immediately next to the surface then reaches saturation, and water vapor changes into liquid droplets on the glass. The droplets can grow, coalesce, and move downward under gravity. If T surface remains above T dew point, the surface does not reach the condition required for condensation.

1. Should I assume a winter case with warm humid indoor air and colder outdoor air?
2. Do you want only the condensation mechanism, or should I also discuss how humidity and window thermal resistance affect it?
3. Should I treat the indoor pressure as approximately atmospheric and focus on the local surface condition?

I would define the physical boundary as the indoor air next to the interior glass surface. Water vapor is the condensing species. For an ordinary building window, I would assume pressure is approximately constant and near atmospheric pressure unless another condition is specified.

The controlling comparison is the surface temperature against the dew point temperature. If T surface is below T dew point, condensation can occur. If T surface is above T dew point, condensation does not occur on that surface.

Heat transfer from the warm room toward the colder exterior lowers the interior glass temperature. The thin layer of indoor air next to the glass is cooled by contact with that surface. As this local air cools at approximately constant pressure, its relative humidity rises. At the dew point, the water vapor partial pressure equals the saturation pressure of water at that temperature. Additional cooling causes some water vapor to leave the gas phase and become liquid.

The phase distribution is therefore humid air in the room and liquid water droplets on the cold interior glass after saturation is reached. During condensation, water releases latent heat. Water can initially appear as small droplets, which grow with continued condensation, coalesce with neighboring droplets, and eventually run downward when gravity overcomes surface retention forces.

The amount and likelihood of condensation depend strongly on indoor moisture content and glass surface temperature. Higher indoor humidity raises the dew point, so condensation can begin at a warmer glass temperature. A window with greater thermal resistance can keep the interior glass surface warmer under the same indoor and outdoor conditions, reducing the chance that the surface falls below the dew point.

I would check the indoor air temperature and relative humidity to determine the dew point, then compare that value with the measured interior glass surface temperature. I would also confirm that the observed droplets are on the interior face because condensation on another face represents a different local temperature and moisture condition. This is ordinary surface condensation, so boiling, critical heat flux, dryout, heat pipes, and vapor chambers are not applicable to this specific question.

Common mistakes are saying that cold air itself creates water, ignoring the dew point, assuming condensation starts whenever the glass feels cold, confusing relative humidity with the actual amount of water vapor, or forgetting that condensation occurs on the surface whose temperature falls below the local dew point.

Lead with the dew point criterion first. Then connect it to heat flowing from the warm interior toward the colder exterior, explain saturation and the vapor to liquid phase change, and finish by stating that the droplets can grow, coalesce, and run down the glass.