A VAV box's reheat coil has been quietly fighting the AHU's cooling coil for the past three years. Neither fault triggers an alarm — the zone temperature is within setpoint, occupants are comfortable, nothing appears broken. But the building is simultaneously heating and cooling the same air stream, burning energy in both directions to achieve a result that could be reached with neither system active. No one noticed, because no one was looking — the trend data existed in the BMS historian the entire time, but reviewing thousands of trend graphs manually for this kind of subtle, silent fault is not a task any human operations team can sustain.
ASHRAE Guideline 36 and Fault Detection and Diagnostics (FDD) analytics solve exactly this problem. Guideline 36 provides the standardised, proven control sequence library that prevents these faults from being introduced in the first place. FDD analytics continuously scan existing trend data — the same data the BMS has been collecting all along — against known fault signatures, surfacing the simultaneous heating/cooling, the stuck damper, the fouled coil that has been silently wasting 8-15% of the building's HVAC energy for years.
Common HVAC Faults: Energy Impact & Detection
| Fault Type | Typical Energy Impact | Detection Method | Time to Detect (Manual) | Time to Detect (AI FDD) | Correction Action |
|---|---|---|---|---|---|
| Simultaneous heating/cooling | 3–6% HVAC energy | Valve position correlation | Months to years (or never) | <24 hours | Sequence reprogramming |
| Stuck/leaking dampers | 2–4% HVAC energy | Position vs. command deviation | Weeks to months | <48 hours | Actuator repair/replacement |
| Degraded coil performance | 2–5% HVAC energy | Approach temperature trending | Months (until capacity loss visible) | 1–2 weeks (trend-based) | Coil cleaning/descaling |
| Economizer faults | 1–3% HVAC energy | OA damper vs. conditions logic | Months to years | <72 hours | Sensor/actuator correction |
| Sensor drift/miscalibration | 1–3% HVAC energy | Cross-sensor statistical comparison | Rarely detected manually | 1–2 weeks (trend-based) | Recalibration/replacement |
Technical Design: FDD & Guideline 36 Architecture
- ASHRAE Guideline 36 sequences: Standardised control logic for VAV boxes, AHUs, and central plants eliminating common programming errors — proven to reduce HVAC energy 15-30% compared to typical non-standardised sequences
- FDD rule engine: Statistical and rule-based detection of 20+ common HVAC fault signatures defined in ASHRAE research — from simultaneous heating/cooling to economizer malfunction to sensor drift
- Simultaneous heating/cooling detection: Correlation logic flagging VAV reheat coils active concurrently with AHU cooling coils — pure energy waste with zero comfort benefit, typically the highest-ROI correction identified
- Economizer fault detection: Verification that free cooling operation occurs during favourable outdoor air conditions per IMD data — a frequent source of missed savings, particularly in India's composite and temperate climate shoulder seasons
- Coil degradation trending: Approach temperature analysis identifying fouled or scaled coils before capacity loss becomes noticeable to occupants — enabling planned cleaning before comfort or capacity impact
- Damper and actuator health monitoring: Position feedback vs. commanded position deviation analysis identifying mechanical failures before they cause zone-level comfort complaints
- Existing BMS trend log integration: FDD analytics typically deployed as a software layer analysing existing trend data via BACnet/IP or database export — no new hardware required in most cases
- India ECBC/BEE context: HVAC dominates 40-60% of Indian commercial building energy; documented FDD implementation and Guideline 36 compliance directly support ECBC Whole Building Performance Method demonstration and BEE star rating improvement
Self-Correcting HVAC: FDD Merges Into Autonomous Remediation
The next evolution of HVAC analytics moves from fault detection to autonomous fault correction — rather than surfacing a flagged issue for a human engineer to investigate and fix, AI-integrated FDD platforms will directly issue corrective control commands within pre-approved safety boundaries (adjusting a stuck-open damper's control signal, rebalancing a simultaneous heating/cooling condition) while escalating only faults requiring physical intervention (actual mechanical repair, coil cleaning) to the maintenance team. This closes the loop between detection and correction, converting today's multi-week manual investigation-and-repair cycle into a real-time, largely autonomous energy optimization process.