Indoor Air Quality Monitoring

People in developed economies spend roughly ninety percent of their lives indoors, so the air inside is the air that actually matters for health - far more than the outdoor air that gets all the attention. Poor indoor air is linked to headaches, fatigue, respiratory irritation and worse, and the pandemic pushed this from a niche facilities concern to a boardroom one: ventilation became a visible proxy for whether a shared space was safe.

The finding that changed how designers think, though, is about cognition. Controlled studies have measured decision-making performance falling as indoor CO2 rises - not because CO2 at these concentrations is toxic, but because elevated CO2 is a reliable marker of under-ventilation, and under-ventilated air carries a soup of other bioeffluents and pollutants. When the air goes stale, people think measurably less well: slower, less strategic, more error-prone. For a knowledge-work building, that is not a wellness nicety - it is productivity leaking away invisibly. This reframes IAQ from a compliance box into one of the clearest business cases in the whole smart-building field: better air is cheaper than the cognitive tax of bad air. It is also deeply relevant in cities with high outdoor particulate loads, where the building envelope and its filtration are the last line of defence for the people inside.

What makes air different from most of the building's other systems is that occupants have almost no way to sense it themselves. You feel when a room is too warm and you see when it is too dark, and you act - but you cannot feel 1200 ppm of CO2 or see PM2.5, and by the time a room smells stuffy the air has been degrading for a while. Air is the one comfort variable that is genuinely invisible, which is precisely why measuring it is so valuable: it converts a hazard people cannot perceive into a number they can. That invisibility also explains why air was historically the most neglected variable - there was no cheap feedback, so it was managed by rule of thumb and worst-case duct sizing. The arrival of low-cost, reliable sensing did not just add a data stream; it made an entire dimension of the indoor environment manageable for the first time, which is why IAQ monitoring has moved so quickly from novelty to expectation.

You do not need a laboratory to monitor air usefully - a handful of signals covers most of what matters. Carbon dioxide (CO2), measured in parts per million, is the workhorse. Outdoor air sits around 420 ppm; a well-ventilated room stays under about 800; above 1000 ppm the air feels stuffy and ventilation is clearly lagging occupancy; well above that, cognition suffers. CO2 is prized because it is a direct proxy for how much fresh air each person is getting.

Fine particulate matter (PM2.5) - particles under 2.5 microns, from traffic, cooking, smoke and outdoor pollution - lodges deep in the lungs and is one of the most health-critical measures, usually reported in micrograms per cubic metre. Volatile organic compounds (VOCs) are gases off-gassed by paints, adhesives, furniture, cleaning products and printers; sensors typically give a relative index rather than a precise concentration, useful for spotting events like a cleaning cycle or a new fit-out. Relative humidity (RH) matters at both ends: too dry (below about 30 percent) irritates eyes and airways, too damp (above about 60 percent) invites mould and dust mites; the 40 to 60 percent band is the healthy target. Some deployments add carbon monoxide, radon or formaldehyde where the risk profile warrants. The art is choosing the few signals that actually drive a decision, calibrating them, and not drowning the twin in noise.

Traditional ventilation is dumb by design: a fixed volume of fresh air on a fixed schedule, sized for a worst-case full room and running whether the space holds fifty people or nobody. That either wastes enormous energy conditioning fresh air no one needs, or under-ventilates when the room is fuller than assumed. Demand-controlled ventilation (DCV) closes that gap by driving the fresh-air rate from a live signal - most commonly CO2, sometimes occupancy counts or a blended IAQ index.

The control loop is exactly the kind you have met throughout this course. An IAQ sensor reports CO2; a controller compares it to a threshold - say 1000 ppm; if the reading climbs above the band it opens dampers or ramps the air handler to bring in more outdoor air; as CO2 falls back it throttles down to save energy. It is a classic feedback loop, often a PID or staged control, closing on air quality instead of temperature. The payoff is real on both sides of the ledger: ventilate a busy room properly and stop over-ventilating an empty one. In a mixed-mode or naturally ventilated building the same logic can prompt occupants - a simple open a window nudge on the app when CO2 climbs. Done well, DCV is one of the clearest wins in smart buildings: measurable energy saved and measurably better air, from the same sensor. Done badly - one poorly placed sensor speaking for a whole floor - it can starve some rooms while flooding others, which is why sensor placement and calibration are not afterthoughts.

Put the pieces together in a single room. It is 1pm; a six-person meeting room is empty and its CO2 reads 480 ppm, close to outdoor air - the overnight purge and the morning's light use have kept it fresh. At 2pm eight people crowd in for a long workshop. With eight sets of lungs in a small sealed box, CO2 climbs fast - past 800 within twenty minutes, through 1000 by half past, heading for 1400 and the stuffy, foggy-headed zone by 3pm if nothing intervenes. On the old fixed schedule, nothing would: the air handler pushes the same volume it always does, sized for an average that this room has just blown past.

With demand-controlled ventilation the story changes. The CO2 sensor crosses the 1000 ppm threshold; the controller ramps the supply of outdoor air to that zone; CO2 stabilises and then falls back toward the band even with the room full. The occupants never notice the intervention - they just keep thinking clearly, which is the entire point. Meanwhile two other things happen quietly. The occupant app can nudge the organiser (this room is filling with CO2 - the system is adding fresh air), and the twin logs the whole episode: a room that regularly spikes might be under-ventilated for its real use, a design insight no static calculation would surface. At 4pm the room empties; CO2 decays; the controller throttles back so the building is not conditioning fresh air for an empty box. That single afternoon shows the whole thesis of the lesson - a signal made visible, a loop that acts on the real load, energy saved at both ends, and a record the twin can learn from - and it also shows the limits, because the same green CO2 reading says nothing about a solvent smell from the whiteboard markers or particulates drifting in from the street. The number is a good servant and a bad master.