Data‐driven flood emulation: Speeding up urban flood predictions by deep convolutional neural networks

Abstract

Computational complexity has been the bottleneck for applying physically based simulations in large urban areas with high spatial resolution for efficient and systematic flooding analyses and risk assessment. To overcome the issue of long computational time and accelerate the prediction process, this paper proposes that the prediction of maximum water depth can be considered an image‐to‐image translation problem in which water depth rasters are generated using the information learned from data instead of by conducting simulations. The proposed data‐driven urban pluvial flood approach is based on a deep convolutional neural network trained using flood simulation data obtained from three catchments and 18 hyetographs. Multiple tests to assess the accuracy and validity of the proposed approach were conducted with both design and real hyetographs. The results show that flood prediction based on neural networks use only 0.5% of the time compared with that of physically based models, with promising accuracy and generalizability. The proposed neural network can also potentially be applied to different but relevant problems, including flood analysis for flood‐safe urban layout planning.

Machine Learning and Urban Drainage Systems: State-of-the-Art Review

“…Meanwhile, the CNN is based on image data and can consider a spatiotemporal distribution for the type of flood-inundation (e.g., flood depth and area, and flood duration) prediction. Kabir et al (2020) [57] proposed a novel CNN model comprising a CNN based on a graphic processing unit to improve model efficiency (e.g., real-time flood forecasting) and performance simultaneously. Real-time-based flood-inundation forecasting models based on advanced measurements (e.g., drones and closed-circuit televisions (CCTVs)) have been proposed recently [60][61][62][63].…”

An ensemble risk assessment model for urban rainstorm disasters based on random forest and deep belief nets: a case study of Nanjing, China

“…For example, Rahmati et al (2019) evaluated the efficiency of a selforganizing map neural network algorithm for URSD and verified the proposed model performed excellently in mapping urban flood hazard. Guo et al (2020) proposed data-driven urban pluvial flood approach based on a deep convolutional neural network and suggested that flood prediction based on NN use only 0.5% of the time compared with that of physically based models. Deep belief nets (DBN) algorithm is a multilayer NN and has a strong classifying ability.…”

A deep learning technique-based data-driven model for accurate and rapid flood prediction

“…To solve the bottleneck problems on complex model construction, long computation time and high cost of the hydrodynamic models, the potential of novel deep learning techniques in capturing and predicting flooding processes have been increasingly explored in recent years to alleviate the burden on physical modelling (Han et al, 2021;Guo et al, 2021;Hou et al, 2021a). The deep learning (DL) 40 methods harbor intelligent learning mechanisms and can extract learning data features from historical knowledge.…”

“…The DL can find the relationships between input and output data with much lower computational cost, in particular with high-performance computers. It has been demonstrated that these DLs have excellent generalization capabilities so that even complex data features (e.g., flood pattern and tendency) can be automatically learned with a high prediction accuracy and computation efficiency 45 (Lecun and Bengio, 1995;Rawat and Wang, 2017;Guo et al, 2021;Yosinski et al, 2014). With proper data provided, the methods can learn the flood patterns through data features and eliminate the analysis of the actual physical processes.…”