Page - 86 - in Intelligent Environments 2019 - Workshop Proceedings of the 15th International Conference on Intelligent Environments

Inputted air temperature 34.7C, radiation 2.03 MJ/m 2 and wind speed 2.6 m/s blowing easterly. In this case, wind is blowing in perpendicular to major street canyons approximately in the same way as for 11.00. The location of overheated spaces is therefore almost similar. However, the amplitude of overheating is less important because the amount of radiation is less important although inputted wind speeds are lower as this does not matter when ventilation is blocked by different building clusters. Figure 9. Air temperature simulations for hour 14.00 within the real environment of Shimbashi (left) and Minamidai (right) at pedestrian level, Tokyo, Japan. Inputted air temperature 34.7C, radiation 2.84 MJ/m 2 and wind speed 3.6 m/s blowing south easterly. The wind is blowing diagonally towards south in this case. This affected the patterns of the spatial distribution of air temperatures. Temperature levels increased due to a high amount of radiation. Temperature levels massively increased within the shallow canopy of the residential district of Minamidai because the design of building blocks is arbitrary and blocking the wind ventilation by creating many recirculation zones although wind speeds were not insignificants. The amplitude of temperature levels within the deep canopy of Shimbashi were less important because high-rise buildings catch winds and accelerate them downdraft and this removed heat and ventilated some open spaces. Conclusion The dynamic approach using an integrated code for energy balance and RANS CFD was tested by using measurements from the COSMO model in Japan. Air temperature patterns and amplitudes fit well between measurements and calculations. This newly developed approach brought more information in terms of the calculation of unstable urban thermal stratification over longer time periods within local boundary layers, especially above and within the upper levels of an urban canopy. This was very clear when the amount of radiation was important which corresponded to hour 10.00 of the simulation. This approach however underestimated air temperature near the ground surface as surface temperatures were quite underestimated. The patterns of temperature distributions were more interesting and informative. The model was tested in real environments and highlighted the fact that deep and/or shallow urban canopies, the urban design of different building blocks and street canyons, wind speeds-direction and the amount of radiation were very important factors for the estimation of overheating within real urban settings. The dynamic approach will bring more information to local urban planning policies aiming to tackle climate change issues such as urban warming. M.Bakkali andY.Ashie /RANSModelling forLocalClimates,EnergyUseandComfortPredictions86