Predicting and characterizing indoor temperatures in residential buildings: Results from a monitoring campaign in Northern Portugal

Empirical data for residential indoor temperature and its determinants have important implications for policymakers in terms of the assessment of thermal comfort, health of occupants and the use for supporting energy demand models. With the purpose of advancing this knowledge, the indoor temperatures of 141 households in the Northern Portugal were measured at a half-hourly basis during the winter of 2013-2014. The observed mean winter daily indoor temperature at the occupied period was 14.9 degrees C for the bedrooms and 16.6 degrees C for the living rooms. The results show that indoor temperatures are significantly below the comfort levels generally accepted, which could be an indication of future potential rebound effects. Results also reinforce the idea that 'cold homes' during winter season are a reality even in the southern European countries. Models for predicting the daily mean bedroom and living room temperature were developed using an enhanced linear regression with panel-corrected standard errors. The results showed that climatic conditions, and especially building characteristics, affect significantly the bedroom and living room's indoor temperatures. (C) 2016 Elsevier B.V. All rights reserved.