Comparison of the performance of local skin temperatures in evaluating overall thermal comfort at normal room temperatures

Local skin temperature is considered to be a potential physiological parameter for thermal comfort evaluation. However, which parts are more suitable for thermal comfort evaluation still needs further investigation. In this study, 20 seated subjects were recruited for thermal comfort and skin temperature monitoring experiments conducted in a climate chamber during winter and summer. The experimental temperatures in winter and summer were 12 °C–24 °C and 24 °C–32 °C, respectively. Thermal comfort evaluation models were developed for the local skin temperature of eight parts via the Fisher discriminant analysis. The performance of the developed models was evaluated using precision and false negative rate (FNR). The results indicated that the skin temperatures of the back of the left hand, left wrist, and left forearm had higher precision for thermal comfort evaluation than those of other parts, whereas the FNR was lower than in the other parts, which were suitable for thermal comfort evaluation. The skin temperature of the left chest had the lowest precision and the highest FNR, making it unsuitable for thermal comfort evaluation. By reasonably selecting body parts, the use of single-point skin temperature to evaluate thermal comfort was better than the use of the mean skin temperature.