Fire

Wildfires in the Mediterranean are expected to become increasingly frequent and intense under climate change and increasing drought conditions. Owing to the growing computational capabilities of modern GPUs, deep learning approaches are increasingly being adopted for wildfire modeling and often outperform conventional methods. This study presents a deep feed-forward neural network, named the Wildfire Neural Network (WNN), designed to generate daily wildfire susceptibility maps. The model integrates meteorological variables, FWI-derived indicators, geomorphological attributes, land-use information, and anthropogenic proxies. WNN was trained and validated using multi-year data from Calabria, southern Italy, a Mediterranean wildfire hotspot. Compared with the operational Fire Weather Index baseline, WNN achieved higher AUROC values and greater spatial information content, with AUROC improvements of approximately 7% to 21% considering the historical configurations.

Wood exposed to fire undergoes both oxygen-independent thermal degradation and oxidative reactions, making the direct transfer of kinetic parameters obtained under inert conditions to fire applications potentially uncertain. This study investigated the effects of atmosphere and heating rate on the thermal degradation and apparent kinetics of Norway spruce (Picea abies (L.) H. Karst.). Simultaneous thermogravimetry and differential scanning calorimetry were performed under nitrogen and synthetic air at heating rates of 10, 25, and 50 °C min⁻¹. Under nitrogen, the principal DTG peak shifted from 346.3 to 364.9 °C with increasing heating rate, while residual mass at 500 °C increased from 21.5 to 27.1%. Under synthetic air, the principal mass-loss event occurred at lower temperatures and was followed by a distinct high-temperature oxidative event accompanied by a strong exothermic DSC response residual mass at 500 °C was 4.6 ± 6.4% and 4.5 ± 3.5% at 10 and 25 °C min⁻¹, respectively. Kissinger apparent activation energies were 263.6 kJ mol⁻¹ under nitrogen and 189.5 kJ mol⁻¹ under synthetic air. Iso-conversional OFW and KAS analyses showed a more consistent kinetic response under nitrogen, whereas synthetic-air data exhibited greater variability and lower regression linearity. The results demonstrate that atmosphere and heating rate substantially influence the thermal response of Norway spruce and provide material-scale experimental constraints for the development and subsequent calibration of solid-phase fire models incorporating both primary wood degradation and oxidative consumption of the resulting solid residue.

The outdoor transformer protection performance of fine water mist, which is widely applied to indoor electrical fire protection, is strongly affected by ambient wind. In this study, the effect of imposed wind speed on water mist distribution was quantified through a full-scale cold spray experiment conducted using a 110 kV transformer in a large wind-tunnel facility. Two nozzle configurations with flow coefficients of 2.0 and 1.7, which corresponded to nominal single-nozzle flow rates of 20 and 17 L/min, respectively, at a terminal operating pressure of 10 MPa, were tested at imposed wind speeds of 0, 4, 6, 8, and 10 m/s. Fifty-eight water collection boxes were used, including 54 boxes on the transformer sidewall and four boxes on the top surface, and the spray duration for each condition was 30 s. The wind-induced redistribution of the mist became evident at 4 m/s and intensified with increasing wind speeds. Water deposition generally decreased on the windward side and shifted toward the downwind region, while the response of the top surface differed because of geometric obstruction and the combined contribution of multiple nozzles. At the same wind speed and operating pressure, the K = 2.0 nozzle configuration maintained a more favorable spray distribution than the K = 1.7 configuration. These full-scale cold spray data clarify the wind sensitivity of outdoor transformer water mist systems and provide an experimental basis for subsequent fire tests and wind-protection design.