Abstract
Generally, the distribution of air temperature is affected by the remnant heat from occupants or the cold draft from openings. So the warm zone always appears at the ceiling level. On the other hand, the cool zone appears at the floor level. The larger the space becomes the more inhomogeneous its temperature distribution becomes. Therefore, it is important to predict such a thermal environment numerically in unsteady state. In this report, we tried to divide a space into three zones (upper, middle and lower) and to apply this proposed three zones model to glass covered spaces by using the Successive Integration Method. And we discussed the improving methods of thermal environment which had an inhomogeneous distribution of temperature.