Database assisted numerical intelligence system (DANIS) for the prediction of aerodynamic wind loads on tall buildings

![CDATA[Numerical prediction of wind loads on tall buildings is well known as a complex Computational Wind Engineering (CWE) problem because turbulent wind flows around tall buildings are highly random, massively separated, and greatly characterized by fluctuating velocity fields. Large Eddy Simulation (LES) is regarded as one of the potential Computational Fluid Dynamics (CFD) techniques to solve this complex problem, but LES is a resource demanding and computationally expensive approach to solve building aerodynamic problems for routine engineering practice. In addition, the clear advantage of computational simulation over wind tunnel simulation is the ability to simulate a building and its surroundings without the need for building physical models, but the approach of time transient simulation such as LES or Detached Eddy Simulation is limited to very small time step and very fine mesh. Another well recognized numerical approach is database approach. This approach has significant advantage to quickly predict wind load but limited to tested specimen in database with similar shape and wind profile. In the present research, an innovative Database Assisted Numerical Intelligence System (DANIS) is currently under development to combine several major numerical techniques including Reynolds Averaged Navier Strokes (RANS), a stochastic Kinematic Simulation (KS), and Radial Basis Functions (RBF) to incorporate experimental results as database and improve the overall efficiency of numerical wind load prediction.]]