Add heat accounting based on supply temperature

[lnagel]

Requesting feedback for PR #64

Available for testing as pre-relase v0.3.0-dev.4

Motivation

Underfloor heating systems need accurate heat accounting to fairly allocate heating time quotas across zones. When multiple zones share a single heat source, simply tracking valve-open duration doesn't reflect actual heat delivered—a zone with its valve open while the boiler is cold receives far less heat than one open during peak supply temperature.

This PR implements supply-temperature-weighted quota tracking: zones accumulate quota proportional to the temperature differential between supply water and room air. This elegantly captures the thermal reality that heat delivery scales with this differential.

Alternatives Considered

1. Full ΔT-based thermal energy calculation (Q = ṁ × c × ΔT)

Requires both supply and return temperature sensors at each zone or manifold, plus a flow meter or flow estimation. Additionally, when the boiler cycles off, the supply-return ΔT collapses while stored thermal energy in the water mass continues heating zones ("coast-down"), requiring complex water mass energy balance calculations to avoid under-counting.

2. Boiler energy meter apportionment

Relies on the boiler reporting a thermal energy counter (e.g., via EMS-ESP), then apportioning energy to zones by their nominal flow ratings. While accurate, this requires specific boiler hardware support and calibrated per-zone flow data that many installations lack.

3. Zone-level supply/return sensor pairs

The most accurate approach, but requires 2× sensors per zone (potentially 16+ sensors for an 8-zone system), significant wiring complexity, and per-zone flow knowledge for proper energy calculation.

4. Flow meter at manifold with zone flow estimation

Requires additional hardware and still needs supply/return temperatures for energy calculation, falling back to the coast-down complexity of option 1.

Why Supply Temperature Weighting

The chosen approach requires only a single additional sensor (manifold supply temperature) and naturally handles boiler cycling: when the boiler fires, supply temperature rises and quota accumulates faster; during coast-down, supply temperature decays and quota accumulation slows proportionally. This provides a practical approximation of heat delivery without the complexity, cost, or failure modes of the alternatives.