Engineering for the Lunar Night
I spent three months reviewing the thermal failure logs from the Shackleton Ridge simulation module. Every system that failed in the dark phase did so not from component fatigue, but from interface degradation — seals that lost elasticity at −160°C, connectors that micro-fractured under diurnal cycling. The hardware was sound. The connections between hardware were not.
Thermal bridges are the real enemy
Conventional habitat design treats insulation as a bulk property — add more layers, reduce heat flux. But on the lunar surface, the dominant loss mechanism is thermal bridging at every penetration point. A single uninsulated bolt connecting an internal truss to an external radiator mount can leak more heat than ten square meters of regolith-shielded dome surface.
The hardware was sound. The connections between hardware were not.
The solution space is narrow but well-defined: every structural penetration must include a thermal break with a conductance below 0.05 W/K. This is achievable with commercially available G-10CR fiberglass standoffs and vacuum-rated PEEK washers. The hard part is the inspection protocol — we now require thermal imaging verification of every bolted interface before pressurization.