Priority-ranked rule roadmap

Implementation order for expanding the public Open-FDD cookbooks. Priorities derive from public literature frequency (ASHRAE GL36 AFDD, Berkeley fault taxonomy, PNNL AIRCx, NIST Cx) and the validated vibe19 catalog.

P0 — validated pandas catalog ✅ (62 rules)

Production Open-FDD SQL registry is 66 rules (sql_rules/registry.yaml). The 62 figure is the pandas oracle catalog floor (see parity matrix).

  • Sensor sweeps: SV-RANGE, SV-FLATLINE, SV-SPIKE, SV-STALE, SV-RATE
  • Control: PID-HUNT-1
  • AHU GL36 FC1–FC15 + AHU-SATDEV / AHU-DUCTHI / AHU-SIMUL
  • Economizer / OA: ECON-1–7, OA-1, OAT-METEO, MECH-OAT-1, VLV-1, DMP-1, CMD-1
  • VAV: VAV-1 through VAV-7, VAV-REHEAT, VAV-AHU-LEAVE
  • RESET-1
  • Plant / CW: CHW-1–4, CHW-NOLOAD-1, CW-APR-1, CW-FAN-1, CW-OPT-1
  • HP-1, WX-1, TRIM-1/3/4, SCHED-1, SCHED-247
  • Framework docs + P0 rule catalog

Next — promote into validated catalog

Rule ID Description
TOWER-1 Cooling tower approach high
CTRL-2 Generic loop hunting
OVR-1 Persistent override
OA-2 DCV minimum OA
PLANT-1 CHW DP reset missing
SP-HIGH / SP-LOW Occupied SP drift
TRIM-2 / KPI-1 / WX-2 Advisory / weather consistency

Later

  • Lead-lag staging anomalies
  • Waterside economizer
  • Heat recovery wheel effectiveness
  • Site-level demand spike vs weather

Documentation milestones

Milestone Status
Canonical taxonomy
Declarative schema
Gap + parity matrices
Prerequisite macros
Doc template
P0 rule catalog (validated metadata)
CI cookbook integrity + fixtures scripts/cookbook_parity_check.py
vibe19 ID sync (Pandas + SQL) ✅ 2026-07-16

Non-goals

  • Vendor-specific rule IDs or threshold bundles
  • Hardcoded device instance numbers or IP addresses
  • Proprietary RCx workbook clones
  • Rules requiring ML models (keep deterministic SQL/Pandas)