Contents lists available at ScienceDirect Fire Safety Journal journal homepage: www.elsevier.com/locate/ firesaf Real-time wildland fire spread modeling using tabulated flame properties Matthieu de Gennaro a,b , Yann Billaud c , Yannick Pizzo b , Savitri Garivait d , Jean-Claude Loraud b , Mahmoud El Hajj a , Bernard Porterie b, ⁎ a NOVELTIS, 153 rue du Lac, Labège, France b Aix Marseille Université, CNRS, IUSTI UMR 7343, Marseille, France c Institut Pprime, CNRS-Université de Poitiers-ENSMA, Poitiers, France d King Mongkut's University of Technology Thonburi, JGSEE, Bangkok, Thailand ARTICLE INFO Keywords: Wildfires Tabulation Genetic algorithm Real-time simulation Front-tracking method ABSTRACT This paper is an extension of previous papers [1,2] on a raster-based fire spread model which combines a network model to represent vegetation distribution on land and a physical model of the heat transfer from burning to unburnt vegetation items, and takes into account local conditions of wind, topography, and vegetation. The physical model, still based on the unsteady energy conservation in every fuel element and detailed local and non-local heat transfer mechanisms (radiation from the flaming zone and embers, surface convection, and radiative cooling from the heated fuel element to the environment), now includes wind-driven convection through the fuel bed. To address the challenge of real-time fire spread simulations, the model is also extended in two ways. First, the Monte Carlo method is used in conjunction with a genetic algorithm to create a database of radiation view factors from the flame to the fuel surface for a wide variety of flame properties and environment conditions. Second, the front-tracking method, drafted in [2], is extended to polydisperse networks and implemented in the new version of the model, called SWIFFT. Finally, the SWIFFT model is validated against data from different fire scenarios, showing it is capable of capturing the trends observed in experiments in terms of rate of spread, and area and shape of the burn, with reduced computational resources. 1. Introduction Currently there are two major approaches to model fire spread: the raster-based approach and the vector-based approach [3]. In the raster-based approach fire spread is treated as a series of cell-to-cell interactions, a set of rules defining the spread mechanism from a cell to its neighbors (see for example [4–9]). The vector-based approach assumes the propagation of the fire front as a continuously expanding polygon and is the basis of the most widely used fire spread models: FARSITE [10], PROMETHEUS [11], and SiroFire [12]. The strengths and weaknesses of both approaches are extensively discussed in [3,13]. One of the main advantages of the raster-based approach is that it is computationally less intensive and is much more suited to hetero- geneous fuel and weather conditions [3]. These features led us to develop a fire spread model based on raster implementation [1,2]. The model combined a monodisperse network (i.e. one in which the fuel elements are close to a single size) to represent vegetation distribution on land with an unsteady physical model of the heat transfer from burning to unburnt fuel elements. The preheating energy-transfer mechanisms considered were: radiation from the flaming zone and embers; surface convection; and radiative cooling from the heated fuel element to the environment. At each time step, overhead flame radiation was calculated by coupling the solid flame model with the Monte Carlo method. In the continuation of these studies, we present here the enhance- ments of the fire spread model that are now being included to improve the scope and validity of the model, and to reduce the computational resources needed to perform simulations. First, in order to improve model predictions of wind-driven fires through highly porous fuels, wind-driven convection inside the fuel bed is included in the model. The second enhancement concerns the calculation of flame radiation during fire spread. Although it provides high accuracy, this calculation requires a large amount of computational resources, which is incom- patible with the operational needs of fire and land management services. In order to run real-time fire spread simulations, radiation calculation is thus performed using a precomputed database of view factors (VF) from the flame to the fuel surface for a wide variety of flame properties and environment conditions. Finally, the front-track- ing method, used to track the fire-front interface by a moving separate grid of lower dimension than the fixed DEM grid [2], is extended to http://dx.doi.org/10.1016/j.firesaf.2017.03.006 Received 24 February 2017; Accepted 15 March 2017 ⁎ Corresponding author. E-mail address: bernard.porterie@univ-amu.fr (B. Porterie). Fire Safety Journal xxx (xxxx) xxx–xxx 0379-7112/ © 2017 Elsevier Ltd. All rights reserved. Please cite this article as: Gennaro, M.D., Fire Safety Journal (2017), http://dx.doi.org/10.1016/j.firesaf.2017.03.006