Search

The fabric air dispersion system (FADS) is a ventilation terminal made of special polymer fabric. The porous structure of the fabric causes complex flow motion. Due to its advantages over the conventional ventilation system, i.e. ducts and diffusers, the FADS has been widely favoured by architects and researchers. In computational fluid dynamics (CFD) simulation the FADS is usually simplified into a free opening with an area equal to all pores and perforations, called the free area (FA) method in this present work. However, the effectiveness of this simplified method has not been validated. The present work took a half cylindrical FADS without orifices as an example and employed the FA method to simulate the airflow properties inside a chamber under isothermal and non-isothermal conditions. The simulated distributions of air velocity and temperature were compared with those by the direct description (DD) method. Meanwhile, the uniformity of air velocity distribution close to the FADS was validated against test data and the flow visualization using the dry ice as a smoking material. Results demonstrate that the FA method is effective and easy to implement, and performs as well as the DD method in predicting the distribution of airflow generated by the FADS without orifices. [ABSTRACT FROM AUTHOR]

The fabric air dispersion system (FADS) has many distinctive features and, when properly designed, can improve indoor thermal environment cost-effectively. In the present work, computational fluid dynamics (CFD) procedures were utilised to investigate the FADS in penetration mode (FADS-PM)-induced flow fields. The mean velocity method (MVel) is proposed to approximately model the supply air flow boundary conditions, which could greatly influence the computational accuracy of overall fluid flow simulations. First, the hypothesis of even velocity distribution at air inlet was validated by flow visualisation and experimental data. Second, the simulation results within the room under isothermal and non-isothermal conditions were compared with those obtained by the direction description (DD) method. Furthermore, the computational resources required were discussed. The results demonstrate that the MVel method performs as well as the DD method in predicting the distribution of airflow generated by the FADS-PM. The MVel method is more easily implemented without considering the physical configurations of the fibre and difficulties in generating grid meshes due to the large scale of length to depth. In addition, the MVel method can greatly reduce the demand for computational resources and save computational time. The results of this research can benefit engineering designers by providing a method that evaluates the overall quality of the FADS-PM-based air-conditioning system more efficiently. [ABSTRACT FROM AUTHOR]