PMV, PPD & Fanger Comfort

Fanger's insight was that the six comfort factors of the last lesson - air temperature, mean radiant temperature, humidity, air speed, metabolic rate and clothing - are not six separate answers but six inputs to one heat-balance equation. Solve that balance and you can predict how a large group of people would, on average, rate the conditions.

That rating is the Predicted Mean Vote (PMV), and it sits on the ASHRAE seven-point thermal sensation scale: -3 cold, -2 cool, -1 slightly cool, 0 neutral, +1 slightly warm, +2 warm, +3 hot. A PMV of 0 means the average person feels neutral - neither warm nor cool. A PMV of +1 means the average person feels slightly warm. The model is called static (or steady-state) because it assumes a person in thermal equilibrium with a fixed environment, exactly as in the climate chamber where it was calibrated. Feed it the six factors and it returns one number - a designed, defensible prediction of how the room will be voted, standardised internationally as ISO 7730 and embedded in ASHRAE Standard 55.

PMV tells you the average vote, but people differ - the same room that is neutral for you is slightly cool for someone else. So Fanger paired PMV with the Predicted Percentage of Dissatisfied (PPD): the fraction of occupants likely to be unhappy at a given PMV.

The striking result is the shape of the PPD curve. It is a U, and its minimum is not zero - at PMV = 0, the most neutral condition possible, PPD is still about 5%. Individual differences mean roughly one person in twenty will be dissatisfied even in a theoretically perfect room. As PMV moves away from zero the curve climbs steeply: at PMV +/-0.5, PPD is about 10%; at +/-1, about 26%; at +/-2, over 75%. This is why ASHRAE 55's static comfort criterion is not 'PMV = 0' but a band - typically -0.5 <= PMV <= +0.5, giving PPD <= 10% - accepting that 10% dissatisfied is the practical best a conditioned building can reach. It is an honest, humbling number: comfort design aims to minimise dissatisfaction, not to eliminate it.

The steepness of the curve also carries a design lesson. Because PPD climbs so fast once PMV leaves the central band, small drifts matter: a room allowed to wander to PMV +1 has more than a quarter of its occupants unhappy, and a swing to +2 loses three-quarters. This is exactly why control matters - a system that holds operative temperature tightly near neutral keeps PPD low, while one that overshoots pays a heavy dissatisfaction penalty for every degree. It is also why comfort complaints in offices cluster: a single warm afternoon can push a whole floor from '10% grumble' to 'half the room emailing facilities'.

Because PMV depends on six variables, its comfort criterion draws out as a zone, not a point. Hold met and clo fixed - say office work at 1.1 met and typical clothing - and plot the remaining factors on a psychrometric chart (operative temperature across, humidity up). The set of conditions that satisfy -0.5 <= PMV <= +0.5 forms a polygon: the comfort zone.

Two features matter for design. First, the zone shifts with clothing: the winter zone (about 1.0 clo) sits lower, roughly 20-24 C operative; the summer zone (about 0.5 clo) sits higher, roughly 23-27 C. Half a clo is worth about 3 C - which is why relaxed summer dress codes save real cooling energy. Second, the zone has humidity limits: too humid and evaporation chokes; too dry and it feels parched and raises static and irritation. ASHRAE 55 defines this zone as the target for mechanically conditioned spaces, and a comfort simulation's job is to show how many occupied hours a design keeps its operative temperature and humidity inside it.

Notice how much of the zone is set by things an architect controls. The horizontal position depends heavily on mean radiant temperature, which glazing, insulation and shading set; a room ringed by warm winter surfaces sits comfortably lower on the air-temperature axis, while one with a sun-struck wall is pushed out the warm side. The zone can also be widened to the right by adding air movement - ASHRAE 55 lets you extend the summer boundary upward when air speed exceeds about 0.2 m/s, which is the formal basis for the fan effect this course returns to. So the comfort zone is not a fixed box the mechanical engineer must hit alone; its shape and position are a design outcome, and simulation is how you show that a better envelope moves the target rather than just chasing it with more cooling.

You do not need to solve Fanger's equation by hand - the CBE Thermal Comfort Tool (Center for the Built Environment, UC Berkeley) does it free in a browser, and it is the fastest way to build intuition. You enter the six inputs - air temperature, mean radiant temperature, air speed, relative humidity, metabolic rate and clothing - and it returns the PMV, the PPD, and a dot placed inside or outside the ASHRAE 55 comfort zone on a psychrometric chart.

The real learning is in moving the inputs. Set a neutral office (24 C, 50% RH, 0.1 m/s, 1.1 met, 0.7 clo) and watch PMV sit near 0. Now raise MRT to 30 C - a sunlit or poorly glazed room - and watch PMV climb into 'warm' even though air temperature has not moved. Drop clothing to 0.5 clo and the dot slides back toward comfort. Add 0.8 m/s of air movement and the comfort zone itself extends to higher temperatures. In ten minutes of sliders you see, quantitatively, everything the last lesson argued qualitatively - and you learn which lever moves comfort most in your case.

PMV/PPD is powerful but has a defined home. It was calibrated on people in steady-state equilibrium with a controlled environment, so it is most valid where those assumptions hold: mechanically conditioned spaces with stable temperatures and occupants who expect a fixed indoor climate - offices, malls, air-conditioned institutional buildings.

Its limits show elsewhere. Field studies repeatedly find that in naturally ventilated buildings, PMV over-predicts discomfort - it calls people 'warm' who report feeling fine, because occupants there adapt (open a window, slow down, accept the season) in ways the static model ignores. It also assumes people cannot control their environment and treats a fixed clo, both untrue in a real, changing building. This is not a flaw so much as a boundary: use PMV for conditioned, steady spaces, and switch to the adaptive model for naturally ventilated and mixed-mode buildings - which is exactly where the next lesson goes. Knowing which model to apply is itself a core comfort-simulation skill.

One more caution about inputs. Because PMV is so sensitive to met and clo, and because those are guesses, a PMV result is only as trustworthy as the assumptions behind it - garbage in, garbage out. A common beginner error is to report a confident PMV to two decimal places while having assumed a single clo value for a whole year. The professional habit is to state your met and clo explicitly, test a plausible range, and present the comfort outcome as a band. Treated that way, PMV is a genuinely useful design instrument; treated as an oracle, it misleads with false precision.