Thermal adaptations and logistic regression analysis of thermal comfort in severe cold area based on two case studies.

• Two thermal comfort investigations were conducted in severe cold area. • The thermal neutral temperatures derived from the linear regression model and logistic model were compared. • The relationship between probability of opening windows and temperature difference was developed. In recent years, the indoor air temperature in residential buildings for space heating rose gradually in the severe cold area of China. Would people feel comfortable in a warmer environment? This paper focuses on the appropriate indoor air temperature in winter. Two field investigations were conducted on the indoor thermal environment and human adaptation in residential buildings with district heating in Harbin during two winter seasons from 2011 to 2014. Ten apartments and their residents were chosen as samples in each investigation. The investigations included measurements of physical environmental parameters as well as subjective questionnaires. The results show that when the indoor air temperature was higher than 24 °C, people would have a sensitive response and preferred a lower temperature. An ordinal logistic regression model was used to examine the result of the linear regression model. The neutral temperature derived from the logistic model is 0.5 °C higher than that from the linear model and the adjusted neutral temperature is 0.8 °C higher than the linear model. Detecting whether people felt "cool" or not and "warm" or not is relatively accurate with the logistic model. It is not sensitive to judge whether people felt "neutral" in low air temperatures. Thermal neutral temperature is not a suitable set-point of indoor air temperature. In the severe cold area of China, the indoor air temperature of 20 °C is suggested as the standard for space heating in residential buildings in terms of comfort and energy saving. [ABSTRACT FROM AUTHOR]