A digital ally in the battle with flames
Can artificial intelligence and algorithms assist Greece to the perpetual – and often ineffective – battle with fires? A start -up company based in London, and a strong Greek « fragrance », has developed an innovative software that, according to its inspirers, could be a valuable tool in the prevention and timely treatment of fires. In fact, the predictions he makes are so detailed that they determine the risk of spreading a fire to individual houses anywhere in the world!
« This is a model that can accurately predict the course and evolution of a fire: where could be manifested and what are its chances of spreading it. None of the existing software provides this feature. The technology we have developed focuses on how a fire can start from a forest area and spread to residential areas, ”says Dr. Savvas Gantonas, co -founder and chief executive of Pinepeak, who is a partnership of four scientists. « It may also predict the time of spreading the fire, it is said to indicate that in one hour the front will extend to this point and in three hours it will be in the other. The prediction is executed within a few seconds, ”said Epaminondas Mastorakos, a professor of engineering at the University of Cambridge, who is a co -founder and scientific adviser to Pinepeak.
Using the software, firefighting and civil protection services could make decisions related to extinguishing and halting a fire or even the preparation of safe evacuation plans of an area. « The model can easily interact with other applications that calculate the time required to move from one point to another. Thus, the best possible decision can be made based on the circumstances, « adds Mastorakos, one of the world’s leading experts in the science of combustion. The development of the model was based on an earlier research on the ignition of aircraft engines. « We have adapted our model to the forest fires, in order to be used by bodies such as fire department, municipalities and regions, and even insurance companies, » explains Cambridge professor. The idea, however, matured after the fire in Eye in 2018. « I grew up in that area. The catastrophe in the eye prompted me to evolve our research. Questions such as « why was one house destroyed, while another was almost untouched? » they remained unanswered. So, we decided to create this tool. «
Flamesight software, as it is called, utilizes a number of data to predict the « behavior » of a fire. « It is based on a natural modeling and uses meteorological and topographic data and data from satellites, utilizing artificial intelligence applications. Its forecasts are so accurate that they reach the scale of an individual plot or home, ”Gadonon explains. « It is the first system to predict how a fire will spread from a forest area within the community. The software takes into account the heterogeneity and diversity of the soil, the weather and the characteristics of the fuel, analyzes the flammability of an area and identifies the differences from road to road, even from building to building (eg, if made of wood or bricks). It then determines the possibility of spreading the fire to a certain point, « adds Gadonas, who was previously a researcher in the Cambridge engineering department.
« Every fire is different. It is, by its nature, a thoughtful phenomenon. The software processes the different natural data and heat released by one fire and quantifies the risk of spreading it or manifesting another fire somewhere else, ”says Mastorakos.
By guide the eye
« We are constantly evolving our technology by studying data from earlier fires. The fire in Mati was the first to model. What we have learned about the interaction of the various flammable materials in a garden, we are now applying them to other areas, « says Gadonas, adding that the software » can be used for both short -term forecasts and medium -term – that is, what can happen in the summer – and 30 years. » The flamesight « could be placed in a business chamber and used in prevention and fire planning » he adds. « We want to work with services such as the Fire Brigade, the Government or the Local Government, in order to make the most of the capabilities of our model. We are open to discussing the possibility of a co -operation that will not burden the Greek taxpayer, « Mastorakos said. But there are other potential software users.
« An insurance company can use it to assess the risk of fire in a particular home before securing it, » Gadonas notes. Pinepeak has already received funding of 350,000 pounds from Genesis Ventures, private investors and the University of Cambridge, and is looking for additional funds to further develop software.
