Physical models of wildfires have been widely applied for evaluating the fire front evolution in order to obtain useful information for fire management. The main drawback in adopting such models for predicting fire propagation concerns the high computational resources that are required, especially when the analysis involves large areas or when multiple scenarios are investigated, e.g. for risk analysis purpose. This is mainly due to the complexity of phenomena involved and the necessity to use multi-dimensional approaches for the reconstruction of the fire front in the landscape. Complete two dimensional analyses can be avoided provided that possible preferential paths can be identified. Once this result is obtained, the analysis is reduced to one dimensional simulations, which can be made faster through the use of reduced models, such as POD models.
In this paper, an entropy generation analysis is proposed as approach to identify preferential paths for fire propagation analysis. The analysis of the effects due to wind and slope as well as the heat transfer mechanisms, including radiation, is performed. The different contributions to the local entropy generation rate are computed in a complex scenario and the areas involving minimum and maximum total amount of entropy generation are considered for evaluating the preferential directions.