Thermal comfort assessment in a residential apartment in Mediterranean climate: a comparative analysis of prediction models
Architectural Science Review, 2026 (AHCI, Scopus)
- Yayın Türü: Makale / Tam Makale
- Basım Tarihi: 2026
- Doi Numarası: 10.1080/00038628.2026.2617411
- Dergi Adı: Architectural Science Review
- Derginin Tarandığı İndeksler: Arts and Humanities Citation Index (AHCI), Scopus, IBZ Online, Avery, Compendex, ICONDA Bibliographic
- Anahtar Kelimeler: comparatıve analysıs, Mediterranean climate, prediction models, residential buildings, Thermal comfort, thermal sensation vote
- Süleyman Demirel Üniversitesi Adresli: Evet
Özet
This study examines the performance of various thermal comfort prediction models in a naturally ventilated apartment located in Izmir, which experiences a Mediterranean climate. Outdoor and indoor temperature and relative humidity levels were monitored throughout April, May, and June. The prediction models evaluated include the Predicted Mean Vote (PMV), Extended PMV (ePMV), New PMV (nPMV), Adaptive PMV (aPMV), and regression models developed from the collected data (rPMV1) and from recent studies (rPMV2, rPMV3, rPMV4). The findings revealed that rPMV1 and nPMV models offered the most accurate predictions. Specifically, rPMV1 demonstrated the lowest bias (−0.01 ± 0.3) and highest accuracy (MSE: 0.10, MAE: 0.26), though this model is inherently specific to the collected dataset. The nPMV model showed strong reliability (bias: 0.01 ± 0.4). Although the traditional PMV model showed a significant correlation with actual Thermal Sensation Votes (TSV), its accuracy was moderate with noticeable bias. The aPMV model, however, performed as the poorest model.