A Long Sequence Urban Block Wind Speed Prediction and Wind-Induced Pedestrian Risk Analysis Method

Extreme weather events are becoming increasingly frequent, resulting in a rise in accidents caused by short-term, strong winds. Wind speeds in urban areas fluctuate more significantly due to the influence of building layouts. This paper proposes a multivariate long-sequence wind speed prediction model specifically designed for urban environments. It incorporates spatial features to capture the patterns of wind speed changes influenced by buildings, thereby enhancing long-term predictions. The experimental results indicate that the average absolute error in predicting the next 12 intervals is only <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$0.0565 \mathrm{m} / \mathrm{s}$</tex>, which validates the model's accuracy. More detailed comparisons were conducted to further demonstrate the superiority of the proposed model. Additionally, the wind forces acting on pedestrians during strong winds are analyzed, and wind speed thresholds that could potentially cause pedestrian distress are calculated. By integrating these prediction results, risk can be assessed and warnings can be issued in real time.