Changing a BMS in a validated or critical lab

Laboratories are among the most energy-hungry buildings a BMS integrator will ever look after, and among the most tightly controlled. The ventilation that makes them expensive is also what keeps people and products safe, so nobody thanks an engineer for a saving that puts a pressure cascade or a fume cupboard at risk. The work is there, but it is done differently: slowly, on paper first, and with the client's safety and quality people in charge of the decisions.

This article is for integrators who service, or want to service, university research labs, pharmaceutical and life-science sites, and similar critical spaces. It covers why the energy is there, who owns which decision, how a change is made on a validated system, and what to retest afterwards. The same discipline applies in data halls, and a short section at the end covers that.

Why labs use so much energy

UK university labs often use three to four times as much energy per square metre as offices, according to audits by the S-Lab programme (S-Lab audit summary) - older data, but the pattern has not changed. The reason is once-through ventilation: air is supplied, conditioned, used once and exhausted, often around the clock and at high air change rates set years ago.

The large North American recommissioning dataset found a median saving of 14% in laboratories, from 28 projects - one of the higher figures for any building type, and a North American one.

Recommissioning savings by building type - Median whole-building energy savings, %
Lodging (hotels)3%
All building types combined (n = 446)6.4%
Data centres (n = 4)11%
Laboratories (n = 28)14%
Public order and safety16%

The lowest and highest segments and three in between - small samples (n = 4) say little on their own.

Laboratories sit near the top of the recommissioning medians. North American figures; small samples say little on their own. Source: Crowe et al. 2020 (LBNL), existing-building commissioning medians by market segment. Flags: North American, Measured. Links: Crowe et al. 2020.

Demand-based ventilation - the evidence, honestly

The best-known results come from the United States. Demand-based control varies air change rates between a low minimum, typically 2 to 4 air changes an hour, and a purge rate when sensors detect contaminants. A study cited in the I2SL guide (2021), covering 1.5 million operating hours in more than 300 lab areas, found the low minimums gave low contaminant levels almost all of the time (I2SL guide). Retrofits under the Smart Labs approach averaged a 58% energy saving across several buildings (US DOE Better Buildings Alliance), and outside-air reduction projects at another US university averaged 27.6% over two years (UNC Green Labs). These are North American results, and the 58% figure came from a summary rather than a report read in full.

UK experience is more cautious. A University of Strathclyde study of a fume cupboard upgrade from constant to variable air volume found real carbon savings, but far less than predicted, because savings depended on how people used the sashes (Oun et al. 2023). The University of Edinburgh's guidance gives rules of thumb - variable volume saving up to about £1,000 a year per cupboard, and closing sashes saving up to 25% (Edinburgh ventilation best practice). And labs dominated by many fume cupboards with high exhaust demand will not see air change savings at all, because the cupboards set the airflow.

Lab ventilation - what demand-based control achieved, and the UK caution - Energy saving, %
Smart Labs retrofits, several lab buildings - average (US)58%
Outside-air reduction projects, two-year average (US)27.6%
Closing fume cupboard sashes - rule of thumb (UK guidance)up to 25%

A UK fume cupboard upgrade from constant to variable volume made real carbon savings, but far less than forecast - sash use decided it. Labs with many fume cupboards will not see air-change savings.

The big US results and the UK caution side by side - offer metered measurement, not forecast savings. Source: US DOE Better Buildings Alliance (UC Irvine Smart Labs); UNC Green Labs; University of Edinburgh ventilation best practice (2024); Oun et al. 2023 (University of Strathclyde). Flags: North American (top two), UK row a rule of thumb, Results depend on sash use.

The practical lesson: sell measurement and verification alongside any ventilation change, and pair variable volume control with sash alarms and feedback to users.

Who owns which decision

In a lab the integrator does not decide how much air is safe. The air change rate is a safety decision, made by the client's environmental health and safety team or biosafety officer, backed by a documented risk assessment. The integrator's job is the controls, the sensing, the sequences and the commissioning that deliver what has been decided - and the evidence that it is being delivered.

In GMP cleanrooms the split is sharper still. EU GMP Annex 1 (2022) gives a guidance minimum of 10 Pa between adjacent rooms of different grades (clause 4.14 of EudraLex Volume 4, Annex 1). The "10-15 Pa" often quoted is industry shorthand, not the Annex 1 wording. Grade A is monitored continuously for particles, a similar system is recommended for Grade B, and monitoring in Grades C and D is set by risk.

The EMS and the BMS are different systems

Many GMP sites run two systems. A separately validated environmental monitoring system (EMS) is the system of record for pressures, temperatures, humidity and particles. The BMS runs the plant under good engineering practice, with targeted qualification of the functions that matter to the product. Practitioners describe this split as common rather than universal (Zamann Pharma), so confirm it on each site.

The split matters commercially. Energy work on the air handling - setting back non-critical areas, reducing air changes in lower grades when rooms are unoccupied - can be possible when the EMS continues to prove compliance and recovery times have been qualified. It still needs formal change control.

The MHRA's GxP data integrity guidance (2018) sets the expectation for records (MHRA guidance): audit trails, controlled user access, time synchronisation, and backups that have actually been restored in a test. A BMS that touches GMP functions needs the same care.

Making a change on a validated system

A change on a GMP BMS follows a path the client's quality team will recognise:

1. Change request. What will change, why, and which rooms and functions it touches. 2. Impact and risk assessment. Which GMP functions are affected, what could go wrong, and how it would be detected. 3. Test script. Written before the work - for a setback, a recovery test showing the room returns to its grade within the qualified time. 4. Execution and evidence. The change made in an agreed window, every step recorded, trends and EMS data kept. 5. Review and handover. Quality assurance signs off, drawings and the controls description are updated, and the change is closed.

Never edit configuration on a GMP function outside an approved change, however small the edit looks. An undocumented change can turn a sound system into a deviation.

What to retest after the work

The same discipline in a data hall

Data centres buy BMS service in a similar way: on risk first and savings second. Uptime Institute's 2025 survey put the global average power usage effectiveness at about 1.54, with little change for six years (Uptime 2025) - so most existing halls still have a cooling gap, much of it controls.

Data centre PUE has barely moved in six years - Power usage effectiveness (total facility power over IT power) - lower is better
UK sites in the first Climate Change Agreement report - base year, mostly 2011 (older UK data)1.93
Global weighted average, 2025 survey (n = 681)1.54
Facilities built in the last five years (same survey)about 1.48
Sites of 20 MW and above (same survey)about 1.44
The floor - every kilowatt-hour reaching the IT (by definition)1.00

The global average has shown little change for six years running - so most existing halls still carry a cooling gap, much of it controls.

PUE has barely moved in six years. The UK figure is older data, from the first Climate Change Agreement report. Source: Uptime Institute Global Data Center Survey 2025 (all three global figures); UK base year from techUK's first Climate Change Agreement report for data centres (the average of the first sites). Flags: Global survey, UK row older data. Links: Uptime Institute 2025 survey, techUK CCA report (archived copy).

One of the first things to fix is airflow. Cold air that bypasses the servers, and hot exhaust that curls back into the inlets, force colder setpoints and harder-working fans. Blanking and containment stop both, and the fans can then slow down - but measure before and after, because many published fan savings are vendor models. Every step is a method of procedure, approved by the client's change board, exactly as in a GMP lab. Data hall cooling inside the envelope covers that work in full.

Shaping the offer

A critical-environment service built on this might include sensor calibration with certificates, alarm and pressure cascade reviews, change control support with written test scripts, fume cupboard retest coordination, recovery testing after setbacks, and metered reporting of any energy measure. Proving savings covers the measurement side, and What the studies really say puts any figure in context.

Related tools and training: the lab demand ventilation estimator (which keeps its estimate marked indicative until the EHS lead's sign-off is recorded), the PID tuning helper, BMS change MOP builder and psychrometrics tools are in tecTools, and the Data centres from the BMS side and Servicing and optimising an existing BMS courses are in tecTraining.

Sources mentioned: S-Lab audits (UK, older data); I2SL guide (2021, North American); US DOE Better Buildings Alliance on Smart Labs (North American, from a summary); UNC Green Labs (North American); University of Strathclyde study (2023); University of Edinburgh ventilation guidance (2024); EudraLex Volume 4, EU GMP Annex 1 (2022); MHRA GxP data integrity guidance (2018); Zamann Pharma (consultancy view); Uptime Institute Global Data Center Survey 2025. Results depend on the lab, its use and its plant.

More articles from 3EC

3EC logo