1. Commissioning is more than switching equipment on
HVAC commissioning verifies the installed system against the owner's operating and performance requirements. A rotating pump, running fan or absence of fault codes is not enough: verify flows, control sequences, operating conditions, alarms and handover.
The process links owner requirements and the basis of design through design review, installation checks, functional testing and operator training. It is not merely a final-day test. https://www.ashrae.org/news/esociety/effectively-applying-the-commissioning-process
This workflow does not replace specialist pressure-testing, electrical safety, refrigerant or manufacturer startup procedures.
2. Define acceptance criteria before testing
A statement such as 'the fan shall work correctly' is not measurable. Specify design flows, temperatures, pressures and tolerances. Define alarm triggers, system response and where operators see them.
Assign designer, contractor, controls team, manufacturer service and operator responsibilities, including witnesses and acceptance authority.
| Test record | Minimum content |
|---|---|
| System identity | Code, location and service area |
| Preconditions | Power, water, cleanliness, access and safety |
| Design target | Flow, temperature, pressure or control requirement |
| Measurement | Instrument, date and actual conditions |
| Acceptance | Specification tolerance and assessment |
| Result | Defect, remedy and retest |
3. Review design and installation documents
Compare selections, as-built drawings and control sequences. A substituted unit may need different flow or sensing. Match pipe sizes, valves, pump curves and fan pressure to installed equipment.
Use consistent labels in the field and controls interface. An ambiguous 'Pump 1' can attach a test to the wrong system. Label valves, diffusers and measurement points.
Trial adjustments cannot replace missing design information. Reducing today's complaint does not establish design performance if targets are unknown.
4. Mechanical prechecks
On water circuits, check flow direction, pump rotation, expansion/air management, strainers, insulation and service valves. Clean before balancing; subsequent filter fouling or trapped air can invalidate readings.
On air systems, check ducts/flexible connections, damper orientation, filters, fan rotation, drains and coils. Compare actual intake/exhaust arrangements with drawings. Forgotten covers and closed dampers can seriously affect performance.
Maintenance access is functional: inaccessible filters, measurement ports or drains make a currently running system difficult to operate reliably.
5. Instruments and stable conditions
Check range, accuracy and calibration for the expected measurement. A wide-range gauge may not resolve a small differential. Wrong valve data can give plausible but incorrect flow readings.
Record active circuits, fan/pump mode, weather and sensing positions. Do not derive capacity from startup transients. Use consistent reference boundaries.
When deriving thermal power, flow and both temperatures must represent the same operating instant. Sensor errors become significant at small ΔT; indirect capacity should not be the sole acceptance evidence.
6. Airside TAB example
Assume total design flow 3,000 m³/h with three terminals requiring 1,200/1,000/800. Measured flows of 1,380/850/770 still total 3,000, but distribution is wrong.
| Terminal | Design | Measured | Deviation |
|---|---|---|---|
| M1 | 1,200 m³/h | 1,380 m³/h | +15% |
| M2 | 1,000 m³/h | 850 m³/h | −15% |
| M3 | 800 m³/h | 770 m³/h | −3.75% |
The example does not establish a general tolerance. Apply the specification and recheck all terminals and total flow after adjustments.
Measure outdoor air separately from total mixed supply. Assess exhaust, building pressure and door behaviour together.
7. Waterside flow and thermal-power example
For 4.0 m³/h water flow and 10 K difference, using approximate density 1,000 kg/m³ and cp=4.18 kJ/(kg·K):
This is heat transfer at the measurement condition. It does not alone prove that a 60 kW unit is deficient; current demand, inlet water and equipment conditions may not require 60 kW. Compare with actual selection conditions.
Record pump flow, differential and speed together. Preserve minimum equipment flow during balancing and part-load tests; do not create an unapproved zero-flow condition by closing all valves.
8. Sensors and actuators
Compare sensors with references. A calibrated room sensor exposed to sun, local heat or a jet may still be unrepresentative: location and accuracy are separate checks.
Observe command, mechanical movement and actual flow effect. A screen showing 100% open does not prove physical travel. Check direction, stroke, linkage and damper failure position.
Verify point scaling and units. Pa/bar or percentage errors can operate suitable equipment incorrectly. Document alarm thresholds with units.
9. Functional test matrix
Define initial condition, action, expected result and observed result. 'Worked' is not a repeatable test record.
| Scenario | Behaviour to verify |
|---|---|
| Normal startup | Correct fan/pump/valve sequence |
| Low load | Stable flow and capacity modulation |
| Sensor fault | Alarm, safe response and recovery |
| Power loss | Controlled stop and suitable restart |
| Dirty filter | Correct pressure alarm and display |
| Standby equipment | Changeover, alarms and duty sequence |
Simulate faults safely within manufacturer limits. Do not bypass protective functions to test. Coordinate fire-alarm interfaces under the approved multidisciplinary scenario.
10. Seasonal and low-load tests
Winter conditions may not prove summer capacity and vice versa. Record deferred seasonal tests and the limits of simulations.
Review early operating trends for continuously high speeds, nearly closed valves and temperature/humidity oscillation. Trends can reveal more than a single handover reading, but first verify their sensors.
11. Defects, training and handover
Assign each defect an owner, deadline and retest. A reported repair is not a passed acceptance test. Sensor or valve changes may require related functional tests to be repeated.
Demonstrate alarms, normal operation, maintenance and emergency actions rather than merely handing over manuals. Deliver as-built drawings, settings, software backups and measurements, explaining why settings were selected.
12. Common errors, checklist and sources
Avoid measuring without targets, accepting only total flow, balancing dirty systems, treating screen values as calibrated references and closing seasonal items without evidence.
Check consistency of design/installation/selections; written criteria and responsibility; suitable calibrated instruments; terminal distribution and outdoor air; full/part-load, fault and standby tests; retests after remedies; and operator training, settings and seasonal plans.
Sources: https://www.ashrae.org/news/esociety/effectively-applying-the-commissioning-process ; https://handbook.ashrae.org/Handbooks/A23/IP/A23_Ch44/a23_ch44_ip.aspx
Examples teach test logic. Tolerances and specialist procedures come from specifications, safety requirements and product documentation.
13. Writing a reproducible scenario
For a dirty-filter alarm, displaying text is insufficient. Verify sensor/threshold, create the condition through an approved method, check recording and operator delivery, then restore normal operation. Preserve safety and product limits throughout.
For standby-pump changeover, simulate failure by a defined safe method; observe start, service continuity, alarm and recovery. Bypassing a protective interlock is not reliable acceptance. Remove and record removal of all temporary overrides after tests, otherwise a correctly tested system can be handed over in a forced state.
14. Measurement uncertainty and acceptance
Record calibration date, range and location. A 0.5 K differential error against a true 5 K difference can affect derived power by about 10%, before flow uncertainty. Take simultaneous readings under stable conditions and avoid claiming high precision from small differences.
The 1,200/1,000/800 versus 1,380/850/770 example has correct total but +15/−15/−3.75% terminal deviations. Total accuracy does not accept every room. Do not derive tolerances from this example; apply contract requirements and remeasure interacting terminals after correction.
15. First season and closeout file
List deferred tests with responsible parties and dates. Distinguish provisional handover from complete scenario verification. Record complaint time, occupancy, weather and mode rather than changing settings after every complaint without a trace.
Closeout includes as-built drawings, selections, balancing, tests, control sequences, alarms and maintenance plans. Train operators in startup/shutdown, faults and energy monitoring. Open items need owners, deadlines and verification evidence. Commissioning proves design objectives through measurement and function, rather than a single start command.