A commercial heat pump can hit every setpoint, raise no alarms, and still burn through a third more electricity than the business case assumed. That is the uncomfortable part of decarbonising heat: the plant does not have to fail visibly to fail commercially. It just runs a little too hot, cycles a little too often, and leans on its backup heater more than anyone realises, and the first hard evidence anyone sees is an energy bill that lands months after commissioning sign-off. By then a full heating season has already been paid for at the wrong efficiency.
The reason is almost always the same. The BMS was set up to control the heat pump but never set up to measure it. Enable signals, schedules and flow-temperature setpoints were configured; energy metering, coefficient-of-performance trending and exception reporting were not. This article is about the second half of that job: what a BMS should actually trend and measure so that a commercial heat pump's performance is proven with data rather than assumed from the absence of complaints. For the connection architecture, sequencing and controls integration that sits underneath it, see our companion guide to commercial heat-pump BMS integration.
A heat pump's efficiency is not a fixed property of the machine. Its coefficient of performance, the ratio of heat delivered to electricity consumed, moves continuously with the temperature lift the compressor is working across. CIBSE AM17:2022, the technical guidance for heat pump installations in large non-domestic buildings above 45 kW thermal, quantifies the sensitivity: each 1K reduction in compressor lift raises COP by roughly 2 to 3 per cent. A unit driven at a 55°C flow when 40°C would have served the building can quietly shed a large share of its rated efficiency without ever tripping a fault. The manufacturer's rated COP, measured under the standard test conditions defined in BS EN 14511:2018, describes the machine on a test rig, not the machine in your plant room on a wet January morning.
That is the monitoring gap. A conventional BMS alarm strategy is built to catch failure: a lockout, a sensor out of range, a unit that will not start. None of those fire when a heat pump is merely inefficient. Poor performance is a trend, not an event, and it is invisible to anything that is not deliberately measuring the ratio of energy out to energy in over time.
Proving performance needs two numbers the BMS often is not collecting: electrical energy in and delivered heat out. The electrical input is straightforward, a meter on the heat pump's own supply rather than a distribution-board total that mixes in pumps, trace heating and everything else. The thermal output is harder and more valuable: a heat meter across the flow and return pipework, combining flow rate with the temperature difference to give delivered kilowatt-hours. With both, the BMS can calculate a live, measured COP and a rolling seasonal figure. Without delivered-heat measurement, the honest fallback is to trend electrical input against outside-air temperature and degree-days, which shows whether consumption is drifting even if it cannot prove absolute efficiency.
The distinction between instantaneous and seasonal performance matters here. A spot COP reading tells you how the unit is doing right now; it is the seasonal figure that decides whether the investment pays back. BS EN 14825:2022 sets the methodology for calculating the Seasonal Coefficient of Performance across a full heating season rather than at a single rated condition, which is exactly why a unit rated at COP 3.8 can deliver a seasonal average nearer 2.8 or 4.2 depending entirely on how it is run and controlled. A BMS that trends measured COP week by week against outside temperature turns that abstract seasonal number into something an FM can actually watch and defend.
Efficiency lives in the temperatures, so the temperature points are where monitoring earns its keep. Flow and return temperatures, and the differential between them, are the headline pair: a heat pump is generally designed around a narrow differential, often in the region of 5°C to 7°C on a low-temperature system, and a persistently poor delta-T is one of the earliest and most reliable signs that something upstream is wrong. Leaving-water temperature trended against outside-air temperature shows whether weather compensation is actually working or whether the unit is being held at a fixed high setpoint inherited from boiler-era logic. Flow rate, buffer-vessel temperatures top and bottom, and outside-air temperature complete the picture, because a COP figure means very little without the conditions it was achieved under.
The single most useful habit is trending leaving-water temperature and delta-T against outside-air temperature on the same axis. If flow temperature is not falling as it warms up outside, weather compensation is not doing its job, and every mild day is being run at an unnecessary efficiency penalty. That pattern is obvious in a trend and completely invisible in a live values page.
We'll assess your controls and provide a detailed quotation with energy savings estimates.
Beyond temperatures, a set of status and counter points expose developing faults long before they become breakdowns. The ones worth trending on every commercial heat pump are:
None of these needs new field hardware in most cases; the points already exist inside the heat pump's controller and are usually available across its BACnet/IP or Modbus interface. Getting them into the BMS and onto a trend log is largely a configuration exercise, and our guide to BACnet versus Modbus for BMS integration covers the practical differences when mapping those registers.
Data only diagnoses if it is logged at a useful resolution. Slow-moving values such as run hours or daily energy totals are fine at longer intervals, but the values that reveal cycling and control instability, compressor starts, flow temperature and delta-T, need short trend intervals, typically a sample every few minutes, or the very behaviour you are hunting for averages itself out of existence. A defrost that lasts three minutes simply will not appear in a fifteen-minute log.
The single most valuable analytical step is weather normalisation. Raw energy consumption always looks worse in a cold month, which tells you nothing. Plotting daily energy or measured COP against outside-air temperature or degree-days strips the weather out and shows the underlying trend, so a heat pump that is genuinely degrading stands out clearly from one that is simply working harder in January. This only works if a commissioning baseline was captured when the system was known to be performing correctly, which is why the baseline should be recorded at handover and stored, not left to be reconstructed from memory a year later when a dispute starts. A building whose controls have drifted over years shows the same symptoms, and our guidance on the signs a BMS needs recommissioning applies directly to heat-pump plant.
Once the data is flowing, a handful of signatures do most of the diagnostic work. A persistently poor delta-T usually means a flow problem, a bypass open when it should be closed, a pump running too fast, or a heat meter reading incorrectly, and it drags COP down because the compressor works across a wider lift than intended. Short cycling, visible as a high starts-per-hour count and short run times, points to a buffer that is undersized or hydraulically bypassed, and it wears compressors out early. Simultaneous heating and cooling, where a reversible unit or a poorly zoned system fights itself, shows up as heating and cooling energy both accumulating in the same period and is pure waste. Excessive backup-heater use is the one that most often explains a shocking bill, and it only becomes visible when immersion or auxiliary runtime is trended separately from the compressor. And a slow, seasonal decline in weather-normalised COP is the fingerprint of a gradual refrigerant loss or fouling that no single alarm will ever catch.
Trending everything is necessary but not sufficient, because nobody watches a hundred trend logs. The performance layer that actually gets used is exception reporting: the BMS quietly comparing live behaviour against the commissioning baseline and against sensible rules, and only raising something when a threshold is crossed. Measured COP below its seasonal expectation for a sustained period, backup-heater hours above a monthly budget, starts per hour over a limit, delta-T outside its band for more than a few hours. A concise dashboard that leads with measured seasonal COP, backup-heater contribution and a weather-normalised energy trend gives an FM or estates manager the three numbers that actually decide whether the plant is earning its business case, with the detailed trends behind them for when something needs investigating.
What good looks like is not complicated: energy in and heat out both metered from day one, the key temperatures and counters trended at a resolution that catches real behaviour, a baseline captured at handover, and a short exception report that turns all of it into a weekly glance rather than a forensic exercise. That is the difference between a heat pump you hope is performing and one you can prove is.
The best time to specify a performance-monitoring layer is before handover, while the heat and electricity meters can still be designed into the pipework and supply rather than retrofitted around them. The second-best time is now, for any commercial heat pump already installed and running blind: adding metering, configuring the trends and setting up exception reporting is a modest piece of work against the cost of a second winter run at the wrong efficiency. If your building has a commercial heat pump and nobody can currently show you its measured seasonal COP, that gap is worth closing before the next heating season. Get in touch to have the monitoring and diagnostics layer scoped for your plant, and see our overview of what a building management system does if you are earlier in the journey.
Specialist BMS installation, commissioning, and maintenance across London and the South East. SafeContractor Approved, BCIA Member.
Our team of building automation specialists is ready to help you optimise your building's performance and efficiency.
Get in Touch
