Heat pumps · Chillers · Sludge drying systems
Knowledge

Water Chiller Commissioning: Sequence, Measurements and Acceptance

Before anything is switched on

Check Acceptance Instrument
Oil heater energised 8–24 h before start Clamp meter on the heater circuit
Oil level in the sight glass Between the marks Visual
Water circuits filled and vented No air at high points Visual
Strainers clean Differential below 10 kPa Gauge
Insulation complete No gaps at valves and flanges Visual
Electrical terminations Torqued, no debris Torque wrench
Insulation resistance Above 1 MΩ Megger
Phase sequence Correct Rotation meter
Sensor readings Match ambient and water temperature Comparison
Safety interlocks Chilled water pump proven before start Simulate

The oil heater is not optional. Refrigerant migrates into the oil when the machine is cold; starting with refrigerant-diluted oil causes foaming, loss of oil pressure and bearing damage within minutes. On a new machine that has been in storage, 24 hours is safer than 8.

Water side first, refrigerant second

Run the chilled water pump and the condenser water pump before the compressor.

Check Target Consequence if wrong
Chilled water flow Within 10% of design Low ΔT, unstable control, freeze risk
Condenser water flow Within 10% of design High condensing pressure
Pressure drop across evaporator Per datasheet Indicates flow and fouling
Pressure drop across condenser Per datasheet Indicates flow and fouling
Flow switch operation Closes before start, opens on stop Nuisance trips or no protection
Water temperatures 7/12 °C chilled, 30/35 °C condenser Design condition

Measure the pressure drop and compare with the datasheet curve. A pressure drop that is 30% low means low flow; 30% high means fouling or a partially closed valve. Both are visible before the compressor runs.

Start-up and loading

  1. Confirm the compressor unloads at start — screw and centrifugal machines normally start unloaded or at low load.
  2. Watch the suction and discharge pressures for the first two minutes; a rapid rise in discharge pressure indicates air in the condenser or a closed valve.
  3. Load in steps, allowing the leaving water temperature to stabilise between steps.
  4. Record current per phase at each step and compare with the datasheet at that condition.
  5. Confirm the slide valve or vane control responds to the leaving water temperature without hunting.
Machine type Start load Loading rate Watch for
Scroll 100% (staged) Step by circuit Short cycling
Screw 25%–40% 5%–10% per minute Oil pressure at low load
Centrifugal 20%–30% Gradual Surge at low load
Magnetic bearing centrifugal 10%–20% Gradual Levitation fault on start

Approach temperatures: the honest performance check

The approach temperature is the difference between the leaving water and the saturated refrigerant temperature in the same vessel. It shows how well the heat exchanger is working, independently of load.

Vessel New machine Acceptable Action needed
Evaporator approach 1.0–2.0 K Up to 3.0 K Above 3 K: fouling or low flow
Condenser approach (water-cooled) 1.0–2.5 K Up to 4.0 K Above 4 K: fouling, low flow, air
Condenser approach (air-cooled) 8–15 K Per datasheet Above: dirty coil, fan fault

Record the approach at commissioning and put it in the log. Every future service visit compares against that baseline, and a rising trend is the earliest warning of fouling.

Capacity verification

Cooling capacity (kW) = flow (m³/h) ÷ 3.6 × 4.186 × ΔT (K)

For 86 m³/h and a 5 K rise, that is 500 kW. Compare with the datasheet at the measured entering condenser water temperature and the measured leaving chilled water temperature.

Reading Expected Diagnosis if outside
Chilled water ΔT 4–6 K Low flow or low load
Condenser water ΔT 4–6 K Flow or fouling
Compressor current Below nameplate FLA Overload or high condensing
COP (measured) Within 5% of datasheet Charge, fouling, or instrument error

Measure COP only at steady state, with stable temperatures for at least 20 minutes, and correct for instrument accuracy. A 5% shortfall is usually instrumentation; a 15% shortfall is a real problem.

Interlocks and safeties to prove

Protection Test Expected
Chilled water flow Stop the pump Compressor stops within 10 s
Low evaporator temperature Lower the setpoint Cut-out before 2 °C
High condensing pressure Throttle condenser water Cut-out at setpoint
Low oil pressure Simulate or observe at start Cut-out within the delay
Motor overload Simulate Trip, no restart without reset
Phase failure Open one phase No start
Anti-freeze heater Power off in cold Heater energised

Test each one, record it, and restore the settings. A protection that has never operated is an assumption, not a safeguard.

Longkang Environmental Equipment Co., Ltd. supplies a commissioning record with the chiller that captures the approach temperatures and the measured capacity, signed by the engineer and countersigned by the client. Phone: +86 400-878-0280 | +86 130-0200-1280

Questions from commissioning engineers

How long does chiller commissioning take?

One day for a single machine with water circuits ready; two to three days for a multiple-machine plant with BMS integration and sequence testing. The oil heater wait adds 8–24 hours before the first start.

Can I commission without the cooling tower?

You can start the chiller on a temporary water supply, but you cannot verify capacity or condensing approach. Commissioning is incomplete until the final heat rejection path is in service.

What if the evaporator approach is 5 K on a new machine?

Check flow first, then water temperature sensors. If both are correct, the vessel may have debris from construction. A new machine should not have a 5 K approach; treat it as a defect, not a characteristic.

Should I record data at full load only?

No. Record at 25%, 50%, 75% and 100% if the load can be staged. Part-load performance is what the building actually sees, and it is where control faults show up.

How do I verify the refrigerant charge?

By subcooling and approach, not by pressure alone. On a factory-charged machine, the charge should not need adjustment at commissioning unless a leak is found.

Is a vibration reading required?

For centrifugal and large screw machines, yes. Take a baseline reading at each bearing at full load and record it. A change of 50% in velocity is a reason to investigate.

Related reading

Frequently asked questions