最新消息 / NEW

2024.08.13 Investigation of integrated artificial intelligence and ensemble forecasting methods on urban inundation

Project Host:Chiang, Yen-ming
Executive unit: Taiwan Agricultural Engineers Society
Subsidy unit: SINOTECH FOUNDATION FOR RESEARCH & DEVELOPMENT OF ENGINEERING SCIENCES & TECHNOLOGIES
Due to climate change and an increase in impermeable surfaces, flooding disasters in metropolitan areas have become more severe in Taiwan. In this study, artificial intelligence methods were used to simulate urban inundation. By integrating a two-dimensional rapid overland flow model with a one-dimensional stormwater sewer model, numerical simulation results under different rainfall scenarios were used as input and reference. An artificial neural network model was developed using AI techniques to predict the status of urban inundation and the ensemble forecasting skill was conducted to provide the final simulations of urban inundation. The results indicated that the neural network model reproduced similar simulation to that of numerical model, with significantly faster simulation speeds. Through the ensemble forecasting skill, the model’s accuracy and stability can be further improved, achieving an error rate of 0.6%