Fault Detection and Diagnostics (FDD) uses automated analysis of BAS trend data to identify equipment malfunctions, control errors, and energy waste before they become costly problems. This guide covers FDD approaches, common HVAC faults, implementation strategies, and organizational workflows for fault response.
FDD continuously analyzes sensor data, setpoints, and equipment status to detect anomalies indicating faults. Unlike traditional alarming which relies on fixed thresholds, FDD uses contextual rules and models to identify subtle operational issues.
| Approach | How It Works | Pros | Cons |
|---|---|---|---|
| Rules-Based | IF-THEN logic comparing points against known fault patterns | Easy to implement, transparent logic | Requires expert knowledge, limited to known faults |
| Model-Based | Statistical models of expected behavior vs actual | Detects novel faults, adapts to building | Requires training data, complex setup |
| AI/ML | Machine learning trained on historical data | Self-learning, pattern recognition | Black box, needs large datasets, can overfit |
Most implementations start with rules-based FDD and add ML capabilities over time.
Description: Heating and cooling systems operate at the same time, wasting energy.
Detection Rule:
IF Heating_Valve > 20% AND Cooling_Valve > 20% FOR > 15 minutes
THEN FAULT: Simultaneous Heating and Cooling
Severity: HIGH (energy waste)
Common Causes: Incorrect deadband configuration, failed changeover logic, sensor in wrong location, competing zone controllers.
Fix: Increase deadband between heating and cooling setpoints to minimum 3F, verify changeover logic, check sensor placement.
Description: Actuator commands change but physical position does not respond.
Detection Rule:
IF ABS(Damper_Command - Damper_Position) > 15% FOR > 10 minutes
AND Damper_Command has changed > 10% in last 30 minutes
THEN FAULT: Stuck Damper/Valve
Severity: MEDIUM
Common Causes: Mechanical binding, failed actuator, disconnected linkage, air in hydraulic lines.
Description: Sensor readings gradually deviate from actual conditions.
Detection Rule:
IF ABS(Zone_Temp_Sensor1 - Zone_Temp_Sensor2) > 3F FOR > 1 hour
AND both sensors in same zone
THEN FAULT: Possible Sensor Drift
Severity: LOW-MEDIUM
Common Causes: Aging thermistor, moisture infiltration, poor calibration, incorrect location.
Description: Outdoor air damper fails to open for free cooling when conditions allow.
Detection Rule:
IF Outdoor_Temp < Return_Air_Temp - 5F
AND Outdoor_Humidity < 65%
AND Economizer_Damper < 30%
AND Cooling is Active
FOR > 30 minutes
THEN FAULT: Economizer Not Engaging
Severity: HIGH (significant energy waste)
Description: Compressor starts and stops too frequently, causing wear and inefficiency.
Detection Rule:
IF Compressor_Starts > 6 per hour
OR Runtime < 3 minutes per cycle
THEN FAULT: Short Cycling
Severity: HIGH (equipment damage risk)
ASHRAE Research Project 1312 established a standardized fault library for air handling units. Key fault categories include: Air-Side Faults (damper stuck, outdoor air fraction incorrect, mixed air temp deviation), Water-Side Faults (valve leaking, coil fouling, pump failure), Control Faults (sensor offset, setpoint error, PID tuning issue), and Equipment Faults (belt slippage, bearing wear, refrigerant leak).
Build fault detection directly in your BAS platform (Niagara, Metasys, etc.).
Niagara Implementation:
1. Create FDD folder in station
2. Add BooleanWritable for each fault condition
3. Create Program object with fault logic
4. Link to alarm extension for notifications
5. Create dashboard showing active faults
Metasys Implementation:
1. Create FDD Logic in LCT
2. Define fault conditions as Boolean objects
3. Route to Alarm Manager
4. Configure notification recipients
5. Add to MUI dashboard
| Platform | Integration | Strengths | Typical Cost |
|---|---|---|---|
| CopperTree Analytics | BACnet, API | Easy setup, good UI | $0.02-0.05/sqft/yr |
| Clockworks Analytics | BACnet, Niagara | Deep ASHRAE 36 rules | $0.03-0.06/sqft/yr |
| SkyFoundry SkySpark | Haystack, BACnet | Flexible, powerful queries | $0.04-0.08/sqft/yr |
| Bueno (Google) | BACnet, API | ML-powered, cloud-native | Custom pricing |
For effective FDD, ensure these minimum points are trended:
| Equipment | Required Points | Trend Interval |
|---|---|---|
| AHU | Supply/return/mixed/outdoor temp, damper positions, valve positions, fan status/speed, filter DP | 5 min |
| VAV | Zone temp, airflow, damper position, reheat valve | 5 min |
| Chiller | Supply/return water temp, flow, kW, status | 1 min |
| Boiler | Supply/return water temp, flow, status, firing rate | 5 min |
| Pump | Status, speed, differential pressure | 5 min |
| Impact | Frequency | Priority | Response Time |
|---|---|---|---|
| High energy waste | Continuous | Critical | Same day |
| Equipment damage risk | Intermittent | High | 24 hours |
| Comfort complaint | Recurring | Medium | 48 hours |
| Minor inefficiency | Rare | Low | Next scheduled maintenance |
Typical FDD Results:
Savings by Building Size:
| Building Size | Annual Energy Waste Found | FDD Cost | Net Savings | Payback |
|---|---|---|---|---|
| 50,000 sqft | $15-25k | $2-3k/yr | $12-22k/yr | Under 1 year |
| 200,000 sqft | $60-120k | $8-12k/yr | $50-110k/yr | Under 1 year |
| 500,000 sqft | $150-300k | $20-40k/yr | $130-260k/yr | Under 1 year |