Factors Affecting Grid-independent Results for Compartment Fire Modelling

Obtaining grid independent results for compartment fires using a computational fluid dynamics (CFD) model is a major challenge, especially when the fire is not prescribed. While simulating a fire scenario using a CFD model, most fire safety engineers use computational cell sizes that can only be supported by their computing resources which may lead to a large error. This paper presents a systematic study to obtain a grid independent result from CFD simulations of an ISO 9705 room fire experiment using the CFD package Fire Dynamics Simulator (FDS) which incorporates a large eddy simulation (LES) methodology along with a mixture fraction combustion model. The experiment involved ignition of two trays of liquid fuel placed in the room and the growth and development of this fire. The study shows that initially as the grid sizes decrease the size of the fire increases and then the fire size starts decreasing to an asymptotic value as the grid sizes decreases further. A discussion is presented on the factors in relation to “goodness ” of the grid resolution such as the changing trend of characteristic fire diameters, the size of computational cells etc