Cleanroom HVAC design and validation, from first calculation to final test.

The air system is the cleanroom

In our essay on cleanroom infrastructure in India, we argued that a cleanroom cannot be talked into passing. This piece is about the system that does most of the passing: the heating, ventilation and air conditioning.

Panels, flooring and seals stop contamination getting in. But particles are generated inside the room every minute, by people, machines and product. Only the air system removes them. It also holds the pressure that decides which way air leaks, and the temperature and humidity that the process and the people need.

That makes cleanroom HVAC design the single largest engineering decision in the project, and usually the largest energy load the plant will carry for twenty years. It is also the system validators spend most of their time testing. Design it from the process, and validation confirms what you already know. Design it from a template, and validation is where you find out.

Start from the process, not the class

Every cleanroom HVAC design begins with a user requirement specification. It records where product is exposed, what each room does, how many people work there and what heat and moisture the equipment releases. The cleanliness class of each room follows from that, not from a catalogue.

Classes are set by ISO 14644-1, and pharmaceutical plants layer GMP Grades A to D on top. Our earlier essay maps one to the other. For HVAC design, one distinction matters most: GMP grades set limits for the room in operation, with people and machines running. Passing an empty room at rest is the easy test. The air system has to be sized for the working shift.

Under the revised Schedule M, binding on all Indian drug manufacturers since 1 January 2026, that distinction now applies to thousands of plants that were never designed around it.

Air changes per hour are a result, not a rule

Cleanrooms use two airflow patterns, and the choice follows the risk to the product.

• Unidirectional airflow sweeps filtered air in parallel streams across the critical zone, such as a Grade A filling line. EU GMP Annex 1 gives a guidance range of 0.36 to 0.54 m/s at the working position, and expects airflow visualisation to prove the stream actually protects the product.
• Non-unidirectional (mixed) airflow serves Grades B to D. Clean air dilutes the particles people and machines generate, and returns carry them away.

In a mixed room, the supply air volume is the largest of three needs: enough to offset heat and moisture gains, enough to dilute particles, and enough to pressurise the room. Tables of typical air changes per hour are a starting point. The regulatory requirement is the outcome: the room must hold its particle limits and recover after disturbance.

Recovery is where the arithmetic shows. In a perfectly mixed room, particle concentration falls exponentially with the air change rate N. The time to reduce concentration by a factor R is:

Annex 1 (2022) asks that at-rest limits are reached about 20 minutes after operations end; ISO 14644-3 tests recovery as the time to fall to 1% of a challenge concentration. Real rooms never mix perfectly, so measured times run longer than the table. That gap, the ventilation effectiveness, is what diffuser placement and low-level returns decide.

The same arithmetic warns against overdesign. A room that clears far faster than 20 minutes is usually moving more air than its risk needs, and paying for it in fan power every hour of its life. Cleanroom HVAC is already the largest energy load in most of these plants.

Pressure cascades decide which way air leaks

Every cleanroom leaks: around doors, through pass boxes, at cable penetrations. The pressure cascade decides the direction. Cleaner rooms are held above their dirtier neighbours, so leakage always runs outward.

The numbers are well settled. Annex 1 (2022), clause 4.14, asks for at least 10 Pa between adjacent rooms of different grades. US FDA guidance asks for 10 to 15 Pa with doors closed, and at least 12.5 Pa from an aseptic room to any unclassified neighbour. ISPE treats 5 Pa as the floor between rooms of the same class.

Good design adds margin to those figures. ISPE’s worked example sets room-to-room setpoints near 15 Pa, an alert at 12.5 Pa and an alarm below 10 Pa, so a room drifting within its normal band never breaches the regulatory floor. Four further decisions shape the cascade:

• Airlock type. A cascade airlock steps pressure down from clean to less clean. A bubble airlock sits above both sides; a sink sits below both. Potent or live products often need negative rooms, which Annex 1 allows when airlocks still protect the critical zone.
• Controlled leakage. Pressure is held by air escaping through designed gaps. Door undercuts, transfer grilles and pressure-stabilising dampers are calculated, not left to the joiner.
• Control strategy. Options run from fixed balance to direct pressure control. ISPE warns against room and AHU flow controls in series, which hunt against each other; the simpler scheme usually holds better.
• Alarm delays. Doors open constantly. Clause 4.16 requires critical differentials to be monitored continuously, with any alarm delay justified in the contamination control strategy, typically a few minutes based on measured door-open times.

Pressure falls one step at a time from the aseptic core to the plant floor; the airlock type sets whether air is pushed out of, or drawn into, the transition.

Filtration, humidity and power: where Indian conditions bite

Filtration

Air passes coarse and fine filters in the air handling unit, then terminal HEPA filters at the ceiling. Grade A zones typically use H14 filters, rated at 99.995% efficiency at the most penetrating particle size under EN 1822. Placing HEPAs at the terminal means a duct leak upstream cannot reach the room. The weak point moves to the filter seal, which is why every installed filter is leak-scanned before handover.

Temperature and humidity

The band comes from the product and the people. Fully gowned operators in a warm room sweat and shed more particles, so comfort is a contamination control, not a courtesy. In India the harder number is moisture. Design on the city’s peak dew point, not only its peak dry bulb: a coil sized for an April afternoon can fail on a humid August evening. Deep cooling with reheat suits most pharma rooms; desiccant wheels take over where very low humidity is needed, as in battery dry rooms.

Power

Grid dips and generator changeovers collapse cascades in seconds. Restart sequences must bring supply fans back before exhaust, so rooms never invert. ISPE recommends keeping environmental monitoring on UPS where power is unreliable, and standby power for Grade A unidirectional zones and isolators, ideally Grade B too. Changeover behaviour belongs in operational qualification, not in the first month of production.

Validation: proving the HVAC does what the design said

HVAC validation follows the same qualification chain as the rest of a regulated plant, with commissioning slotted between design and installation checks. Each stage produces evidence the next one relies on.

Most of the measurable work sits in operational qualification. ISO 14644-3:2019 sets out the methods: air pressure difference, airflow, airflow direction and visualisation, recovery, temperature, humidity, installed filter leakage and containment leak tests among them. Classification itself is a particle count under ISO 14644-1, at a number of locations set by the room’s floor area.

“A room qualified in a dry February has not yet met its hardest month.”

Two habits make the difference on site. Have an accredited agency that reports to the client run the tests, not the contractor. And in India, extend performance qualification monitoring across a monsoon where the programme allows.

After handover: monitoring and requalification

Qualification proves the room on the day. Monitoring proves it every day after.

ISO 14644-2:2015 asks each site to set its monitoring plan from a risk assessment. It caps the interval between particle concentration tests at 6 months for ISO 5 and cleaner, and 12 months for less stringent classes. Annex 1 fixes requalification of Grade A and B areas at no more than 6 months, and Grades C and D at 12 months. Critical pressure differentials are monitored continuously, not sampled.

The calendar is only half of it. Any change to the system, whether a replaced filter, a rebalanced damper or a new machine adding heat, should pass through change control and trigger the tests it affects. A rising pressure drop across filters, a fan drifting up its speed range or a cascade that needs constant retuning are early signs. A good building management system trends them before an inspector does.

Where cleanroom HVAC projects in India go wrong

The failures are rarely exotic. Most trace back to a decision made early and never revisited.

• Treating it as air conditioning. Comfort HVAC manages temperature for people; cleanroom HVAC manages contamination for product. A contract that only says “AC” will not deliver a cascade, a recovery time or a qualification file.
• Designing on dry-bulb alone. Coils that hold temperature but not moisture pass in winter and fail in the monsoon.
• Leaving leakage to chance. A cascade calculated on paper fails when panel joints, ceiling penetrations and door gaps leak more, or less, than assumed. Leakage paths should be drawn, sealed and tested.
• Reusing old plant in Schedule M upgrades. Existing air handlers often lack the air volume, filter stages or static pressure the new zoning needs, and ceiling voids may be too shallow for HEPA housings and ducts.
• Never testing a power cut. If changeover is not simulated in operational qualification, the first real outage becomes the test.
• A file that does not match the room. Tests run but not witnessed, as-builts never updated, excursions never closed. Inspectors read records before they read gauges.

Design it to be tested

The best cleanroom HVAC designs are written with the validation protocol open on the next screen. Every air volume has a calculation behind it, every pressure step has margin, every leakage path is drawn, and every failure mode has a test planned before the first duct is hung.

That is the same discipline we apply to every project. Our industrial facility in Andhra Pradesh routed airflow, drainage and MEP for its clean-air zones as one 3D model, using the BIM process we have written about. We are not a specialist sterile-suite contractor, and we say so. We are an engineering-led builder organising ourselves for exactly this kind of measurable work.