Predicting the flame characteristics and rate of spread in fires propagating in a bed of Pinus pinaster using Artificial Neural Networks

Chetehouna, Khaled; Tabach, Eddy El; Bouazaoui, Loubna; Gascoin, Nicolas

doi:10.1016/j.psep.2015.06.010

Cited by 25 publications

(9 citation statements)

References 39 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…However, the model is trained on semi-empirical models rather than experimental data. Chetehouna et al [47] used an artificial neural network to estimate the RoS, flame height, and flame tilt angle in pine needle beds. In future works, we plan to use high-resolution datasets such as the one we have collected to train machine learning models that can be used to parameterize small-scale dynamics.…”

Section: Discussionmentioning

confidence: 99%

Fine-Scale Fire Spread in Pine Straw

Sagel

Speer

Pokswinski

et al. 2021

Fire

View full text Add to dashboard Cite

Most wildland and prescribed fire spread occurs through ground fuels, and the rate of spread (RoS) in such environments is often summarized with empirical models that assume uniform environmental conditions and produce a unique RoS. On the other hand, representing the effects of local, small-scale variations of fuel and wind experienced in the field is challenging and, for landscape-scale models, impractical. Moreover, the level of uncertainty associated with characterizing RoS and flame dynamics in the presence of turbulent flow demonstrates the need for further understanding of fire dynamics at small scales in realistic settings. This work describes adapted computer vision techniques used to form fine-scale measurements of the spatially and temporally varying RoS in a natural setting. These algorithms are applied to infrared and visible images of a small-scale prescribed burn of a quasi-homogeneous pine needle bed under stationary wind conditions. A large number of distinct fire front displacements are then used statistically to analyze the fire spread. We find that the fine-scale forward RoS is characterized by an exponential distribution, suggesting a model for fire spread as a random process at this scale.

show abstract

Section: Discussionmentioning

confidence: 99%

Fine-Scale Fire Spread in Pine Straw

Sagel

Speer

Pokswinski

et al. 2021

Fire

View full text Add to dashboard Cite

show abstract

“…Transition rules for the CA were determined by the ELM trained with data from historical fires, as well as vegetation, topographic, and meteorological data. Likewise, Chetehouna et al (2015) used ANNs to predict fire behavior, including rate of spread and flame height and angle. In contrast, Subramanian and Crowley (2017) formulated the problem of fire spread prediction as a Markov decision process in which they proposed solutions based on both a classic RL algorithm and a deep RL algorithm; the authors found that the DL approach improved on the traditional approach when tested on two large fires in Alberta, Canada.…”

Section: Fire Spread and Growthmentioning

confidence: 99%

A review of machine learning applications in wildfire science and management

et al. 2020

View full text Add to dashboard Cite

Artiﬁcial intelligence has been applied in wildﬁre science and management since the 1990s, with early applications including neural networks and expert systems. Since then the ﬁeld has rapidly progressed congruently with the wide adoption of machine learning (ML) methods in the environmental sciences. Here, we present a scoping review of ML applications in wildﬁre science and management. Our overall objective is to improve awareness of ML methods among wildﬁre researchers and managers, as well as illustrate the diverse and challenging range of problems in wildﬁre science available to ML data scientists. To that end, we ﬁrst present an overview of popular ML approaches used in wildﬁre science to date, and then review the use of ML in wildﬁre science as broadly categorized into six problem domains, including: 1) fuels characterization, ﬁre detection, and mapping; 2) ﬁre weather and climate change; 3) ﬁre occurrence, susceptibility, and risk; 4) ﬁre behavior prediction; 5) ﬁre eﬀects; and 6) ﬁre management. Furthermore, we discuss the advantages and limitations of various ML approaches relating to data size, computational requirements, generalizability, and interpretability, as well as identify opportunities for future advances in the science and management of wildﬁres within a data science context. In total, we identfied 300 relevant publications up to the end of 2019, where the most frequently used ML methods across problem domains included random forests, MaxEnt, artiﬁcial neural networks, decision trees, support vector machines, and genetic algorithms. As such, there exists opportunities to apply more current ML methods — including deep learning and agent based learning — in the wildﬁre sciences, especially in instances involving very large multivariate datasets. We must recognize, however, that despite the ability of ML methods to learn on their own, expertise in wildﬁre science is necessary to ensure realistic modelling of ﬁre processes across multiple scales, while the complexity of some ML methods, such as deep learning, requires a dedicated and sophisticated knowledge of their application. Finally, we stress that the wildﬁre research and management communities play an active role in providing relevant, high quality, and freely available wildﬁre data for use by practitioners of ML methods.

show abstract

“…Therefore, a large number of different approaches have been developed for modeling. There are quite a lot of studies in this regard [37][38][39][40] . 41) , particularly about the opportunity of ML implementation in the scenario framework for the application of forest and land fire DRR technology.…”

Section: Fire Behaviour Predictionmentioning

confidence: 99%

Machine Learning Application in Response to Disaster Risk Reduction of Forest and Peatland Fire

Riza¹,

Santoso

Kristijono

et al. 2023

Majalah Ilmiah Pengkajian Industri

View full text Add to dashboard Cite

Peat forest is a natural swamp ecosystem containing buried biomass from biomass deposits originating from past tropical swamp vegetation that have not been decomposed. In the dry season, this accumulated biomass called peat, which was originally submerged in water, is exposed to the surface and prone to fire. Once it burns, smouldering peat fires consume biomass 15 times larger than open flame. Peat smouldering fires are very difficult to extinguish. These will continuously occur for weeks to months. The most recommended effort by experts and practitioners of peat smouldering fires is to prevent them before they occur with the strategy: 'detect early, locate the fire, deliver the most appropriate technology'. Monitoring methods and early detection of forest and land fires or 'wildfire' have been highly developed and applied in Indonesia, for example monitoring with hotspot data, FWI (Fire Weather Index), and FDRS (Fire Danger Rating System). These 'physical simulator' based methods have some weaknesses (1) the accuracy is very low; (2) having predictive biases in areas with different ecosystem characteristics; (3) it is often difficult to implement because it requires a large and complex number of 'expert rules'; and (4) involving many variables, complex, and heterogeneous data formats. Slowly but surely this method has been replaced by the Machine Learning (ML) method as it is developing in Europe, China, India, Japan, North America and Australia. What about the potential application of ML in the forest and land fires, especially smouldering peat fires in Indonesia? This paper tries to answer this question through the following points: State of the art of machine learning in science and management of forest and land fire Outlook on technology of Disaster Risk Reduction (DRR) in BPPT in the field of forest and land fire and the opportunities of ML implementation Impact based forecasting and machine learning for DRR of forest and land fire This paper recommends a conceptual design: Impact-Based Learning for DRR of Forest, Land Fire and Peat Smouldering

show abstract

Predicting the flame characteristics and rate of spread in fires propagating in a bed of Pinus pinaster using Artificial Neural Networks

Cited by 25 publications

References 39 publications

Fine-Scale Fire Spread in Pine Straw

Fine-Scale Fire Spread in Pine Straw

A review of machine learning applications in wildfire science and management

Machine Learning Application in Response to Disaster Risk Reduction of Forest and Peatland Fire

Contact Info

Product

Resources

About