There are numerous applications in architecture field that use CFD simulations either for indoor or outdoor spaces including urban design, ventilation systems, wind load analysis and thermal comfort analysis. The air modeling technologies that have been used so far (wind tunnel measurements, CFD) are usually embedded in later design stages and none of them can meet the needs for time saving in order to integrate this kind of simulation in earlier stages. So, as the factor of time efficiency is obviously very important, Fast Fluid Dynamics (FFD) that introduced by Foster and Metaxas for computer graphics are starting to gain increased interest. (Foster and Metaxas, 1996). Stam (1999) suggested a different approach for fluid simulations that was developed mainly for the gaming industry implementing a faster visually satisfying way but no so accurate in order to simulate wind behavior. As Stam mentioned his fluid solver suffers from “numerical dissipation” (Stam, 1999) and it is not appropriate to be used for engineering problems. This disadvantage is noticed from Zuo and Chen (2007, 2009 and 2010) as well, pointing out that the results suffer from lack of turbulent flows but after an investigation and comparison of results between CFD and FFD in three different cases (1. flow in a lid-driven cavity 2. flow in a plane channel and 3. flow in a ventilated room) the FFD method was validated and proved that can produce reasonably accurate simulations much faster than other computationally-hungry techniques. Thus, FFD managed to make early analysis and exploration of complex phenomena possible and gave the designer a long awaited handy tool, opening new horizons to the whole design approach.
In initial paper presented by Stam, the algorithmic approach for the solver in a 2D environment was explained as well as its easy portability to a three dimensional one. Mike Ash implemented the 3D FFD simulation code in C programming (‘real time 3d fluid simulation’) based on Stam’s paper (Stam, 2003) and later it was Chronis (2010) who first mapped that into processing.js environment. Chronis et al. (2011) used FFD simulation as form finding method and in combination with GA tried to optimize the smoothness of a curved surface to minimize the tensions introduced by air on it (Fig_16). A paper that was based on Stam’s 2D FFD simulation is written by Sheby (2011) and investigates the wind comfort criteria in outdoor spaces using simultaneously Agent based model in order to predict people’s movement behavior in these areas. Another paper that uses Chronis et al. 3D FFD simulation was implemented by Karagkouni (2012) and explores the apertures configurations of a facade in order to achieve better indoor ventilation providing human comfort. All these papers implemented FFD in combination with GA in order to find the optimum solutions in this complex problem of air.
In initial paper presented by Stam, the algorithmic approach for the solver in a 2D environment was explained as well as its easy portability to a three dimensional one. Mike Ash implemented the 3D FFD simulation code in C programming (‘real time 3d fluid simulation’) based on Stam’s paper (Stam, 2003) and later it was Chronis (2010) who first mapped that into processing.js environment. Chronis et al. (2011) used FFD simulation as form finding method and in combination with GA tried to optimize the smoothness of a curved surface to minimize the tensions introduced by air on it (Fig_16). A paper that was based on Stam’s 2D FFD simulation is written by Sheby (2011) and investigates the wind comfort criteria in outdoor spaces using simultaneously Agent based model in order to predict people’s movement behavior in these areas. Another paper that uses Chronis et al. 3D FFD simulation was implemented by Karagkouni (2012) and explores the apertures configurations of a facade in order to achieve better indoor ventilation providing human comfort. All these papers implemented FFD in combination with GA in order to find the optimum solutions in this complex problem of air.
Fig_16