With the development of computer numerical simulation technology, large eddy simulation (LES) has become an important research method for wind environment of microscale (single building), local scale (small scale, building group) and even mesoscale (medium scale) in urban canopy. The subgrid scale model of LES is one of the key factors of numerical simulation accuracy because of the high Reynolds number in urban wind field. With the expansion of simulation scale in urban wind environment, solid wall of building (group) requires a large number of grids. Therefore, a new approach to deal with solid wall of building and boundary grid needs to be developed. This paper studied multi-scale drag term model of buildings (group) to replace the influence of solid buildings (group) on fluid flow, which simplifies the solid obstacle under urban canopy to porous media and simplifies the flow past obstacle into the flow through different resistance sources. For the numerical simulation of wind field, the applicability of universal subgrid-scale model (USM) had been studied. The wind environment of local scale and mesoscale was studied by using the drag term model of different scales. Adding the influence of thermal environment on the wind field, the local scale drag term model was applied to city planning of Xixian New District which could propose more scientific and reasonable planning suggestions for ventilation corridors. General research results are as follows:

1) Simulation accuracies of classic Smagorinsky model, dynamic Smagorinsky model and universal subgrid-scale model (USM) had been studied through flow past circular cylinder with different low, medium and high Reynolds numbers. The results showed that USM did not need additional coefficient, nor depended on empirical coefficient and empirical equation, which could be applied to complex turbulent with fewer grids and higher accuracy of numerical simulation. Dynamic mixing length of USM based on turbulent kinetic energy, dissipation rate and similar scale hypothesis dynamics, which could reflect the spatial and temporal distribution of subgrid-scale fluctuation and improve the accuracy of LES.

2) To clarify the controlling factors of urban canopies, a series of single high-rise buildings were simulated under fully developed turbulence at a realistic scale by large-eddy simulation (LES). The results show that the recirculation area and recirculation length were proportional to the frontal area scale (width/height) and inlet velocity, but was inversely proportional to the length of the building and angle of wind direction.The existing method to estimate recirculation range was improved through summarizing the law of different building volumes. The maximum error of recirculation area formula was only 7.5% (near the ground) and the maximum error of recirculation length formula was only 6.7%. The recirculation range formulas laid a theoretical foundation to study the drag term parameter model of single building. According to the low accuracy of current drag term parameterization scheme, a new drag term model was proposed which divided the energy loss into three parts from conservation of energy: the momentum loss caused by the obstruction in the frontal direction of the obstacle, the loss of surface friction resistance caused by the surface of the barrier and the surface, and the dissipation loss caused by the stickiness of the fluid in the recirculation region behind the barrier. The drag term length was found to be proportional to the recirculation area, frontal area and surface area; and inversely proportional to the porosity and height, but was not closely related to floor area. The maximum error is only 11.7% and the average error is only 5.1%.

3) Drag term parameterization scheme of single building had been expanded to local scale. The aligned array configuration, staggered array configuration and aligned array with different height configuration had been used to verify the solid wall results and drag term model results using Belcher’s, Santiago’s and our drag term model. The results show that Belcher and Santiago models are more suitable for low and sparse building group but not applicable for high-rise building. However, the effective range of frontal area rate in this paper was 0-1, which could accurately reflect the block ratio of the building to the inlet velocity. Particularly, there is no recirculation range behind the front three rows of buildings in staggered array configuration, but the building after fourth rows surrounded by the recirculation range. Therefore, in urban planning, try to keep buildings with staggered array configuration before front three rows, and then reserve some space (park, green area, lake and so on) to ensure flow field recovery. Pure building drag term model could be expanded to the mixing model with building and vegetation. The results have shown that local scale drag term model was easy to be expanded to different surface types. The drag term parameterization of this paper was applied to Xixian New District. Through analyzing the average velocity at 1.5m, 5m and 12m, two first-level corridors, two second-level corridors and some three-level corridors were suggested. Decreasing the height of building reasonably is good for strengthening the air exchange rate in the city.

4) Influence of thermal environment on wind field had been studied with Xixian New District based on the local scale drag term model under the effect of the temperature field. Firstly, the average temperature of typical climate days in 2014 had been calculated with improved CTTC model. Three typical urban surface types had been selected to confirm the initial temperature field of Xixian New District. The result showed that local scale drag term model with temperature field could reflect the influence of local thermodynamic circulation on wind field accurately. There was only one penetrating wind corridor at pedestrian height and the wind corridor become more obvious when the number of obstacles gradually decreases as the altitude increases. As the obstacle effect in Northeast and central of Fengdong new area was too big, three wind corridors could not be connected thoroughly near the ground. It is suggested to decrease the height of buildings at low wind corridor or increase the green area, which could form cold source to promote thermodynamic circulation.