FIELDMIND LABCOMPUTATIONAL R&D RU
T1Thermal systems2026

Numerical proof-of-concept

T1: Phase-Guided Graphene Thermal Interface

A claim-bounded thermal evidence chain tests phase-guided redistribution, controls, tensor direction, action-normalised policies, and a final topology comparison.

T1.10 confirmatory thermal outcome and hotspot-state localisation charts
T1.10 joint action-state topology comparison from the public evidence ledger.

01

Research question

Does topology retain a measurable thermal effect after cumulative controller action and observable tensor state are closely matched?

02

Method and experiment series

Ten stages progress from balance and passive baselines to fixed-budget redistribution, controls, corrective replays, action normalisation, and a full-production source-versus-anti-hotspot confirmation.

Series: T1.1–T1.10 · 88 figures

03

Key results

  • T1.9 is a negative result: after action normalisation, the least adverse hotspot policy remains 0.098% worse off-centre and 0.134% worse for the dual source.
  • In T1.10 the source-aligned topology improves impedance by about 0.148% versus reference, while the anti-hotspot topology degrades it by about 0.255%.
  • The final thermal separation is about 0.403%, with hotspot-state localisation separation ≈ 0.671.

04

What is not established

The effect is small, controlled, and model-specific. It does not establish graphene kinetics, actuator energy, microscopic trajectory identity, laboratory thermal conductivity, manufacturability, or product performance.

This work is a numerical proof-of-concept. It establishes model-level behaviour under the stated assumptions and does not replace independent laboratory validation.