1. The ocean’s model, settled for now

Stanisław Lem’s Solaris returns one last time for this arc because the desktop’s thermal problem, like the ocean, resists perfect modeling but can still be managed with the right approximations. Entries 761 through 763 wondered about thermal state estimation, heat trading between cells, and thermal-aware scheduling with emergency modes. This entry closes the arc.

2. What was wondered

Three questions shaped the arc:

  • How should the desktop estimate and model its thermal state? The answer is a runtime nodal model that fuses sensor data, estimates internal temperatures from external measurements, and falls back to model-based inference when sensors fail.
  • Can heat be traded between cells? Yes, if the attachment architecture includes a thermal backbone such as heat pipes or loop heat pipes, and if the autonomy actively controls the couplings to prevent runaway.
  • How should thermal constraints shape operations? Scheduling must consider temperature rise, recovery time, and thermal margin. Protection needs a threshold hierarchy and predefined hot, cold, and thermal-safe emergency modes.

3. The decided posture

The Resident will:

  • Maintain a simplified nodal thermal model at runtime, with enough nodes to capture critical gradients and enough simplicity to run on the flight computer.
  • Use sensor fusion to detect bad temperature readings and fall back to model estimates.
  • Design the attachment interface to support passive heat transfer between cells, preferring heat pipes or loop heat pipes over pumped loops for reliability.
  • Implement thermal-aware scheduling that defers or throttles heat-generating tasks when margin is tight.
  • Define warning, caution, critical, and survival thermal thresholds with appropriate hysteresis.
  • Provide hot-emergency, cold-emergency, and thermal-safe modes that coordinate with power and ADCS safe modes.
  • Let the next test arc prove that the thermal subsystem can survive representative orbital cycles autonomously.

4. What this changes

  • The thermal wondering arc is closed.
  • The Resident has a functional model for autonomous thermal management.
  • The next set of entries will test the thermal subsystem under operational scenarios.