Comparative analysis of PMV Models accuracy implemented in the ISO 7730:2005 and ASHRAE 55:2023

The Predicted Mean Vote (PMV) model, developed by Fanger, is included in the ISO 7730:2005. It has a low prediction accuracy, especially outside thermal neutrality. Several other formulations have been proposed to compensate for its limitations. One is the PMV calculated in accordance with the ASHRAE 55:2023 (PMV CE ) standard. The ISO 7730:2005 and ASHRAE 55:2023 are the most widely referenced thermal comfort standards worldwide, hence, we compared the accuracy of the two PMV models, against 49 245 thermal sensation votes collected in buildings. We found that overall both models only predict Thermal Sensation Vote (TSV) correctly only one-third of the time. The PMV CE has a higher bias than PMV in predicting thermal sensation in the subset of data with relative air speed ≥ 0.2 m/s. Both models have an accuracy lower or equal to 5% when predicting either ‘warm’, ‘hot’, or ‘cold’. This value is lower than random guessing (14%), meaning they are misleading users. A PMV value higher than 0.5 or lower than -0.5 only indicates that people’s thermal sensation is on the ‘warm’ or ‘cold’ side, respectively. Hence, we suggest reducing the applicability limit of both models within the range ∣ PMV ∣ ≤ 0 . 5 . Finally, the assumption that individuals who feel ‘slightly warm’ or ‘slightly cool’ can be considered thermally neutral leads to under-counting thermal discomfort. In the ASHRAE Global Thermal Comfort Database II, 68% of participants who were ‘slightly warm’ wanted to be ‘cooler’ and 54% of participants who were ‘slightly cool’ wanted to be ‘warmer’. • We compared the accuracy of the PMV models in ISO 7730:2005 and ASHRAE 55:2023. • Both models have low and similar prediction accuracy (PMV = 32% – PMV CE = 34%). • The PMV CE has a higher bias than PMV in predicting thermal sensation. • We recommend limiting the applicability range of the models to ∣ PMV ∣ ≤ 0 . 5 . • It is misleading to classify slightly warm/cool individuals as thermally comfortable.

Authors

Institutions

Publication Details

Journal
Building and Environment
Published
2025-03-06
DOI
https://doi.org/10.1016/j.buildenv.2025.112766
Citations
19
Primary Topic
Building Energy and Comfort Optimization
Type
article
Field-Weighted Citation Impact
19.12
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

Comparative analysis of PMV Models accuracy implemented in the ISO 7730:2005 and ASHRAE 55:2023

Stefano Schiavon, Federico Tartarini
19 citations
Building and Environment
Building Energy and Comfort Optimization
19.12
article

Comparative analysis of PMV Models accuracy implemented in the ISO 7730:2005 and ASHRAE 55:2023

Stefano Schiavon, Federico Tartarini
article en
19 citations

Abstract

The Predicted Mean Vote (PMV) model, developed by Fanger, is included in the ISO 7730:2005. It has a low prediction accuracy, especially outside thermal neutrality. Several other formulations have been proposed to compensate for its limitations. One is the PMV calculated in accordance with the ASHRAE 55:2023 (PMV CE ) standard. The ISO 7730:2005 and ASHRAE 55:2023 are the most widely referenced thermal comfort standards worldwide, hence, we compared the accuracy of the two PMV models, against 49 245 thermal sensation votes collected in buildings. We found that overall both models only predict Thermal Sensation Vote (TSV) correctly only one-third of the time. The PMV CE has a higher bias than PMV in predicting thermal sensation in the subset of data with relative air speed ≥ 0.2 m/s. Both models have an accuracy lower or equal to 5% when predicting either ‘warm’, ‘hot’, or ‘cold’. This value is lower than random guessing (14%), meaning they are misleading users. A PMV value higher than 0.5 or lower than -0.5 only indicates that people’s thermal sensation is on the ‘warm’ or ‘cold’ side, respectively. Hence, we suggest reducing the applicability limit of both models within the range ∣ PMV ∣ ≤ 0 . 5 . Finally, the assumption that individuals who feel ‘slightly warm’ or ‘slightly cool’ can be considered thermally neutral leads to under-counting thermal discomfort. In the ASHRAE Global Thermal Comfort Database II, 68% of participants who were ‘slightly warm’ wanted to be ‘cooler’ and 54% of participants who were ‘slightly cool’ wanted to be ‘warmer’. • We compared the accuracy of the PMV models in ISO 7730:2005 and ASHRAE 55:2023. • Both models have low and similar prediction accuracy (PMV = 32% – PMV CE = 34%). • The PMV CE has a higher bias than PMV in predicting thermal sensation. • We recommend limiting the applicability range of the models to ∣ PMV ∣ ≤ 0 . 5 . • It is misleading to classify slightly warm/cool individuals as thermally comfortable.

Building and EnvironmentVol. 275
The University of Sydney (AU), University of California, Berkeley (US)
Openalex Percentile: Top 0%
Building Energy and Comfort Optimization
19.12
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.