関東の住宅における窓開閉・冷暖房使用の相対的行動モデルの提案

抄録

 In daily life, we conduct the various occupants’ behaviours for thermal comfort and health. The stochastic models of occupants’ behaviours were analysed by some previous studies. Especially, the models of window open in previous studies were analysed by Gaussian function, polynomial logit and logistic regression linked to linear function in FR mode (Free Running: without heating & cooling being used mode). However, Gaussian function has the symmetrical curve, and the regression coefficient of polynomial logit need many number of digits. And, the model of window open in FR mode was not compared with air conditioner use. Moreover, it is important to understand the relation between the opportunities of each behaviours because these affect each behaviours. From these, we need to analyse the relative proportions of each behaviour in the integrated single stochastic model with the concept of figure 1. In order to integrate the models of the window open/closed, cooling (CL) and heating (HT) use, we analysed the occupants’ behaviour model using the logistic regression method with the data collected in Japanese dwellings. We have measured the air temperature, relative humidity and globe temperature in 120 dwellings, and conducted occupants’ behaviour survey during 4 years. From this series of survey, we have collected 36,154 responces all together. The integrated model was calculated by the following equation.

 P_W = P_FR×P_WFR ={1 – (P_HT + P_CL)}×P_WFR

 where P_W is proportion of window open in mixed mode (considering FR, CL and HT), P_FR is proportion of FR mode, P_HT is proportion of heating use, P_CL is proportion of cooling use and P_WFR is proportion of window open in FR mode. This integrated model showed the relative proportion of each behaviours can be expressed as follow.

 P_HT + P_CL + P_W + P_WCFR = 1

 where P_WCFR is the proportion of window closed for neither heating nor cooling use in mixed mode. The major findings are as follows: 1) the logistic regression equations of window open/closed, cooling and heating use were integrated into a single occupants’ behaviour model as a function of outdoor air temperature. This model represents the relative proportion of each behaivours: window open/closed, heating on/off and cooling on/off. 2) In this model, the maximum proportion of window open was 0.47 at 26.0 °C of outdoor air temperature. The proportion of window open decreases when outdoor air temperature is either lower or higher than 26.0 °C. The proportion is calculated by the increasing the cooling use. 3) The regression coefficient of this model was simpler than the polynomial logit. The regression curve of window open model is asymmetric. 4) When the outdoor air temperature was 28.3 °C, the proportions of either window open or cooling use are 0.44. If the outdoor air temperature is lower than 28.3 °C, the proportion of window open become higher than cooling use, and vice-versa. From these results, it can be said that this integrated model is effective to know the combined effect of occupants’ behaviours on the using of window, mechanical heating and cooling. Moreover, this integrated model is useful to implement in the building thermal simulation for predicting the energy use in buildings.