To evaluate the influence of the rooftop surface cooling on the indoor heat load, we conducted a dynamic heat load calculation by the thermal circuit network method (Ishida et al., 1987). Furthermore, to evaluate the influence of the rooftop surface cooling on the atmospheric heat load, we conducted an unsteady heat conduction calculation of a one­dimensional multi-mass system on the roof using the backward relaxation method, which made it possible to predict more detailed data such as a cross-sectional temperature distribution. For the evaporation and the reflection of solar radiation on building outer surface, we used new sol-air temperature taking evaporation into consideration. Details of the SAT* is given in the following:

Qa= Qr-Qv-Qie Eq. (1)

Since Qr= Qs-Qi, Eq. (1) can be rewritten as follows:

A.(<30/<3z) = Qs-Qi-Qv — Qie Eq. (2)

Here,

Qv= ac(0s-0a) Eq. (3)

Qs= (1-p)Js Eq. (4)

equation (fs=a*0s+b) of 0S,

Qie = wTaw(fs-fa) = a* wTa, w(0s-0a) + w*l*aw(a*0a +b-fa) Eq. (6)

Organizing Equations (2)~(6), the following equation can be obtained:

-A.(<30/<3z) = at[(0a+0e*) — 0s] Eq. (7)

Here, however, at and 0e* will be

at = (1-K*CC)* e*ar+ ac+ a*w*l*aw Eq. (8)

0e* = [(1-p)Js — (1-K*CC)(1- Br) *eaTa4 — wTaw(b-fa)] / at Eq. (9)

In this study, (0a+0e*) is handled as the sol-air temperature taking evaporation into consideration (SAT*, hereafter). Fig. 5 shows the results for which SAT* was calculated using the hottest day the maximum air-conditioning load was generated (August 5) from the standard weather data of Tokyo. SAT* dynamically changes depending on the solar reflectance (p, hereafter) rather than the evaporation rate (w, hereafter). If the value is identical, the SAT* with a value of p that is greater than that of w takes precedence. Here, based on the concept of above SAT*, we propose the equivalent solar reflectance (p*, hereafter) to evaluate the effects of evaporation and the reflection of solar radiation. If equivalent outside air temperatures with evaporation in mind are equal, we considered the

effects of evaporation and solar reflection to be equal. In the following, we conducted a numerical analysis with p* as the main parameter.

Fig. 6, 7 shows the floor plan of the building and the rooftop slab structure in the indoor heat load calculation. For the subject, we envisioned an office building; the plan for the room was made based on standard office simulations of AIJ (Architectural Institute of Japan). The rooftop surface temperature can vary greatly depending not only on the surface finish but also the insulated state of the rooftop slab. For this reason, a calculation was made for the case with no insulation (thermal resistance: 0.51m2K/W) on the rooftop slab.

Then, after changing the thickness of the insulation material in five stages, a calculation was conducted for the case of internal insulation (thermal resistance: 0.87~5.88m2K/W) on the rooftop slab. Tab. 2 shows various input conditions to calculate the indoor heat load. Air­conditioning was set to be on between 8:00 and 18:00 and the fixed temperatures for air­conditioning and heating were set at 26°C and 22°C respectively. Inside the building, the heat generation of 25W/m2 from lighting and equipment was established. The convection heat-transfer rate on the indoor side and outdoor side was 11.1 and 25.0W/m2K respectively.