[ Solution ]
Chiller and Cooling System Monitoring
In a hotel, a hospital or a shopping centre, cooling is the largest line on the summer electricity bill. Yet on most sites no number is ever produced that says how efficiently the chiller ran that day.
The chiller is running. But is it running well?
Your building management system tells you the chiller’s state: whether it is enabled, what the water temperature is, whether an alarm is active. That is useful, and it is not efficiency.
Efficiency comes down to one question: for every kilowatt of electricity spent, how many kilowatts of cooling are produced? That ratio is the COP, and calculating it needs more than electrical power. It needs the cooling load — which means chilled water flow and the difference between flow and return temperature.
On most sites those three measurements do not exist together. So nobody knows whether the chiller is running well or badly. They only know the bill grows in summer.
COP cannot be compared without the weather
A chiller’s efficiency depends on ambient conditions. The same machine performs noticeably better on a 24 °C day than on a 30 °C one. So “COP was 5.2 last month and 4.8 this month” says nothing on its own — the weather may simply have changed.
For this we pull hourly meteorological data for the site’s location: dry bulb and relative humidity. On water-cooled plant the decisive figure is the wet bulb, derived from those two, because it sets the lowest temperature the tower can possibly reach.
That separates a change in COP that came from the weather from one that came from the plant. Monitoring that cannot make the distinction raises an “efficiency has dropped” alarm every summer, and everyone learns to ignore it.
Machine COP or system COP?
This distinction lands directly on the bill.
Machine COP counts only the electricity the chiller itself draws. System COP counts everything spent to produce that cooling: chiller, primary and secondary pumps, tower fans.
The gap between them is not small. If the chiller is performing well but the pumps run at fixed speed, or the tower fans run harder than they need to, machine COP looks healthy and the bill stays high anyway.
We show both, separately. The panel makes it obvious which item is drawing what, and that is where the search for improvement starts.
What we measure, and from where
Measurement plate
- Electrical power
- kW
- Chiller, pumps and tower measured separately
- Chilled water
- °C · m³/h
- Flow, return and volumetric flow rate
- Condenser circuit
- °C · m³/h
- Inlet, outlet and flow rate
- Cooling load
- kW
- Derived · flow × ΔT
- COP
- kWth/kWe
- Derived · machine and system reported apart
- Ambient
- °C · %RH
- Hourly meteorology · wet bulb derived
- Tower approach
- °C
- Tower outlet − wet bulb
Where a measurement point cannot be added, the datasheet curve is used as the basis and that is labelled explicitly. Measured and derived are never presented as the same thing.
The largest saving is usually in the sequencing
On a plant with more than one machine, the biggest gain rarely comes from improving a single chiller. It comes from which machine starts first.
Even two units of the same model diverge with age and fouling. If the sequencing order is never updated to reflect that, the plant runs the worse machine first, every day, indefinitely.
We compare them across the load bands they share: which machine is more efficient in which band, how running hours are distributed, and how far the part-load curve sits below the datasheet reference. Sequencing stops being a setting somebody chose years ago and becomes a decision backed by measurement.
The tower sets the ceiling
When efficiency drops, the chiller is the first suspect. The most common cause sits outside the machine.
A cooling tower’s outlet water can, in theory, be cooled to the wet bulb temperature. The gap between them is the approach. As the approach widens, condenser water returns warmer than it should, and the chiller draws more electricity to produce exactly the same cooling.
A widening approach usually points to the tower’s fill, its water distribution, or its fan control. You cannot find it by looking at the chiller. You cannot find it at all without measuring the tower — which is why the approach is one of the values we track continuously rather than calculate on request.
And the condenser side
The condenser’s inlet-to-outlet difference, together with its flow rate, shows whether heat is genuinely being rejected. A fouled condenser, or a flow rate that has quietly fallen, pulls COP down without ever raising an alarm — nothing trips, nothing warns, the bill simply grows.
We follow both the tower and the condenser on their own pages, because these are the findings worth having: cheap to correct, and the correction holds.
Common questions
Our BMS already monitors the chillers. What would change?
A BMS manages state: start, stop, hold a setpoint, raise an alarm. It does not calculate efficiency, because that would require measuring the cooling load — flow and ΔT — and ratioing it against electrical power. Most BMS installations have no flow meter. Where we can take data from your BMS we will; we add the measurement that is missing.
The manufacturer already publishes a COP. Why measure it?
A datasheet COP holds at the design condition, on a new machine. Yours runs at a particular part load, at a particular condenser temperature, with however many years on it. The difference between the two can run to tens of percent. We put the measured COP alongside the datasheet curve and show the gap.
Does installation require work on the chiller itself?
Nothing is opened up. The measurements needed sit on the pipework and in the electrical panel: a flow meter, temperature sensors and a power analyser. Where existing analysers or BMS points are available they are read instead. If the chiller's own control panel exposes data, that can be taken too.
We have an absorption machine. Can it be compared with an electric chiller?
It cannot, and we do not. An absorption machine's thermal COP is calculated against heat input; an electric chiller's against electricity. They are not on the same scale, and reading them side by side misleads. The panel reports them separately and carries that warning above the screen.
Does this support ISO 50001 or ESOS?
It supports both, without being either. For ISO 50001 the hard part is normally continuous measurement of significant energy uses and a baseline you can defend — that is what this provides. For ESOS the assessment itself must be signed off by an approved lead assessor, but the twelve months of data behind it, and the progress reporting the scheme now expects against an action plan, come from the same measurement layer. The Phase 4 compliance deadline is 5 December 2027.
Let us measure what is happening on your site.
In a one-hour call we look at your existing setup and set out exactly which measurement points are needed and what you would be able to see.