ABSTRACT
This chapter introduces the integration of the genetic algorithm (GA) method and adjoint/artificial neural network (ANN) method. Fast fluid dynamics (FFD) and a new numerical algorithm, semi-Lagrangian scheme with the PISO algorithm (SLPISO), have also been evaluated for their potential to reduce the computing time of computational-fluid-dynamics (CFD) simulations. The chapter describes effort to accelerate CFD simulations by integration of the GA method and adjoint method, integration of the GA method and ANN method, and the use of faster numerical algorithms. It introduces the combination of the GA method with an artificial neural network in the inverse design of an indoor environment. The chapter discusses the feasibility of the GA-based method. It implements four FFD models, non-incremental pressure-correction scheme with the semi-Lagrangian (SL) scheme (NIPC-SL), NIPC, standard incremental pressure-correction (SIPC), and rotational incremental pressure-correction (RIPC), in Open source Field Operation And Manipulation (OpenFOAM) and validated them for predicting steady-state and transient flow in indoor environments.
