In recent years, new fire loads dominated by power banks have caused multiple fire incidents in transportation hubs, posing severe challenges to fire safety. This study uses experimental testing and machine learning regression analysis to explore the heat release rate (HRR) characteristics and influencing factors of box-type power banks under fire conditions. A 1 MW calorimeter was used to conduct four sets of experiments involving a total of 15 box-type power banks, measuring the HRR and analyzing its correlation with oxygen consumption, carbon dioxide generation, smoke temperature, and the fire growth rate. Based on the experimental data, HRR prediction models were constructed using decision tree regression (DT), K-nearest neighbor regression (KNN), and linear regression (LR). The results indicate that the DT model performs best in HRR prediction (MAE = 0.4889, MSE = 0.7414, RMSE = 0.8571, R² = 0.9991), effectively capturing the nonlinear variation patterns in the HRR. The correlation analysis and regression analysis conducted in this study provide crucial data support for fire combustion characteristics research, fire risk assessment, and fire safety optimization. Furthermore, the findings provide crucial data support for research on fire combustion characteristics and data-driven fire risk assessment, which may serve as a foundation for future AI-based real-time fire detection applications.

Excessive pilot mental workload during crosswind approach and landing affects pilots’ ability to maintain a stable approach and landing and may cause the aircraft to run off the runway. This paper reports a flight simulator study involving 24 cadet pilots with real flight experience. We used 38-channel noninvasive functional near-infrared spectroscopy (fNIRS) to collect the participates’ brain data. Subjective workload ratings and fNIRS data containing three hemoglobin signals, that is, changes in the concentration of oxyhemoglobin (ΔOxyHb), changes in the concentration of deoxyhemoglobin (ΔDeoxyHb), and changes in the concentration of total hemoglobin (ΔTotalHb), were collected from 48 approach and landing scenarios with and without crosswind. A total of 4,218 effective connectivity (EC) features were extracted from three hemoglobin signals across all sampling channels using Granger causality (GC). Statistical analysis was conducted on the subjective ratings and EC features, and EC features with absolute correlation coefficients greater than 0.2 were selected as inputs for the models. Combining TabNet and decision tree (DT), a stacking ensemble learning model (TabNet-DT) was established as a classifier for evaluating pilot mental workload and was compared with TabNet and DT. The results suggested that brain EC can be used to differentiate various levels of pilot mental workload and that ΔTotalHb was the most sensitive to pilot mental workload. Compared with the other models, TabNet-DT demonstrated superior performance, achieving 93.19% accuracy after selecting the combination of features of different hemoglobin signals. The findings from this study can contribute to improving flight safety during approach and landing.