When a fire occurs in a street canyon, smoke recirculation is the most harmful factor to human beings inside the canyon, while the wind condition is an essential factor determining if the smoke is recirculated. This paper focuses on the wind direction’s effect on buoyancy-driven fire smoke dispersion in a street canyon, which is innovative research since the effect of wind direction has not been reported before. In this study, an ideal street canyon model with a height–width ratio of 1 was established, and both the wind velocity and wind direction were changed to search for the critical point at which smoke recirculation occurs. The results show that with an increase in the wind direction angle (the angle of wind towards the direction of the street width), the smoke recirculation could be distinguished into three regimes, i.e., the “fully re-circulation stage”, the “semi re-circulation stage”, and the “non-recirculation stage”. The critical recirculation velocity was increased with the increase in the wind direction angle, and new models regarding the critical wind velocity and the Froude number were proposed for different wind direction conditions.
Industrial visual anomaly detection remains challenging because practical inspection systems must achieve high detection accuracy while operating under highly imbalanced data, diverse defect patterns, limited computational resources, and increasing demands for interpretability. This work aims to develop a lightweight yet effective and explainable anomaly detection framework for industrial images in settings where a limited number of labeled anomalous samples are available. We propose a dual-branch feature-based supervised ensemble method that integrates complementary representations: a PCA branch to capture linear global structure and a scattering branch to model multi-scale textures. A heterogeneous pool of classical learners (SVM, RF, ET, XGBoost, and LightGBM) is trained on each feature branch, and stable probability outputs are obtained via stratified K-fold out-of-fold training, probability calibration, and a quantile-based threshold search. Decision-level fusion is then performed by stacking, where logistic regression, XGBoost, and LightGBM serve as meta-learners over the out-of-fold probabilities of the selected top-K base learners. Experiments on two public benchmarks (MVTec AD and BTAD) show that the proposed method substantially improves the best PCA-based single model, achieving relative F1_score gains of approximately 31% (MVTec AD) and 26% (BTAD), with maximum AUC values of about 0.91 and 0.96, respectively, under comparable inference complexity. Overall, the results demonstrate that combining high-quality handcrafted features with supervised ensemble fusion provides a practical and interpretable alternative/complement to heavier deep models for resource-constrained industrial anomaly detection, and future work will explore more category-adaptive decision strategies to further enhance robustness on challenging classes.
This study investigated smoke dispersion inside a street canyon in a series of numerical simulations. The building height and street width as well as the cross-wind velocity were changed during the simulation, and the smoke recirculation behavior inside the canyon is presented and discussed. The results show that the smoke recirculation behavior could be distinguished into two different stages, i.e., the “fully recirculation stage” and “semi recirculation stage”, which is strongly determined by the canyon aspect ratio (the building height divided by street width). It was found that the critical wind velocity at which the smoke recirculation would take place was almost constant for an ideal street canyon with an aspect ratio of 1; however, this velocity was decreased with increasing building height or decreasing street width, indicating a much more dangerous circumstance when the aspect ratio is greater. Finally, a new piecewise function is proposed for the critical smoke recirculation velocity for all cases, which can provide some theoretical basis for building designs and emergency rescue for human beings inside the street canyon.