Two-dimensional convolutional neural network outperforms other machine learning architectures for water depth surrogate modeling

Yan, Xiaohui; Mohammadian, Abdolmajid; Ao, Ruigui; Liu, Jianwei; Yang, Na

doi:10.1016/j.jhydrol.2022.128812

Cited by 17 publications

(7 citation statements)

References 43 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…CNN utilize convolutional layers to efficiently extract spatial feature information from input data, necessitating fewer parameters relative to fully connected neural networks 21 , 33 , 34 . Furthermore, an Artificial Neural Network (ANN) was also employed for the reconstruction of datasets related to the maximum daily storm surge.…”

Section: Methodsmentioning

confidence: 99%

A dataset of storm surge reconstructions in the Western North Pacific using CNN

Dang,

Feng,

et al. 2024

Sci Data

View full text Add to dashboard Cite

The relatively short duration of available tide gauge records poses challenges for conducting comprehensive statistical analyses of storm surges in the Western North Pacific. To address this issue, we employ a convolutional neural network model to reconstruct the maximum daily storm surge at 160 tide gauges from 1900 to 2010 in the Western North Pacific. The reconstructed dataset serves multiple purposes. Firstly, it facilitates the identification of regions where notable changes in the storm surges have occurred in the past. Additionally, the dataset enables long-term analyses of the storm surge climate, offering insights into historical patterns and variations. Furthermore, it provides a solid foundation for conducting robust extreme value analyses. To ensure accessibility, the data are publicly available through a repository, allowing for easy access and utilization by the broader scientific community and the general public. Overall, our research contributes to the field of oceanography by providing a dataset that aids in understanding the historical storm surge dynamics in the Western North Pacific region.

show abstract

Section: Methodsmentioning

confidence: 99%

A dataset of storm surge reconstructions in the Western North Pacific using CNN

Dang,

Feng,

et al. 2024

Sci Data

View full text Add to dashboard Cite

show abstract

“…These methods have proven to be highly effective in reducing the computational cost for tasks such as optimization and sensitivity analysis [1,23]. Several techniques have been utilized for constructing surrogate models, including radial basis (RB) [24][25][26][27], least squares support vector machines (LS-SVMs) [18,28], and artificial neural networks (ANNs) [9,15,[29][30][31][32][33][34].…”

Section: Introductionmentioning

confidence: 99%

“…Conversely, if the data are insufficient for training the surrogate, it may not achieve proper generalization, hindering the identification of optimal solutions. Thus, there is ongoing interest in investigating techniques that simplify the implementation of surrogate models while minimizing computational time and maximizing model accuracy [1,15,33,35]. In this regard, the development of new strategies and EAs that are well-suited for exploring solutions in simulated solution spaces generated with surrogate models could significantly accelerate the processes of calibrating complex distributed models, extending its range of application in water resources research and management.…”

Section: Introductionmentioning

confidence: 99%

Surrogate-Assisted Evolutionary Algorithm for the Calibration of Distributed Hydrological Models Based on Two-Dimensional Shallow Water Equations

Farfán-Durán,

Heidari,

Dhaene

et al. 2024

Water

View full text Add to dashboard Cite

Distributed hydrological models based on shallow water equations have gained popularity in recent years for the simulation of storm events, due to their robust and physically based routing of surface runoff through the whole catchment, including hill slopes and water streams. However, significant challenges arise in their calibration due to their relatively high computational cost and the extensive parameter space. This study presents a surrogate-assisted evolutionary algorithm (SA-EA) for the calibration of a distributed hydrological model based on 2D shallow water equations. A surrogate model is used to reduce the computational cost of the calibration process by creating a simulation of the solution space, while an evolutionary algorithm guides the search for suitable parameter sets within the simulated space. The proposed methodology is evaluated in four rainfall events located in the northwest of Spain: one synthetic storm and three real storms in the Mandeo River basin. The results show that the SA-EA accelerates convergence and obtains superior fit values when compared to a conventional global calibration technique, reducing the execution time by up to six times and achieving between 98% and 100% accuracy in identifying behavioral parameter sets after four generations of the SA-EA. The proposed methodology offers an efficient solution for the calibration of complex hydrological models, delivering improved computational efficiency and robust performance.

show abstract

“…Data-driven approaches based on modern machine learning (ML) algorithms for water-related problems have become a topic of significant research due to both academic and practical interests [34][35][36][37][38][39][40] and have recently been employed to complement CFD models. The existing studies either used ML algorithms to improve hydrodynamic models or directly developed surrogates for the numerical models.…”

Section: Introductionmentioning

confidence: 99%

Fast Prediction of Solute Concentration Field in Rotationally Influenced Fluids Using a Parameter-Based Field Reconstruction Convolutional Neural Network

Yan

Mohammadian

Yu³

et al. 2023

Water

Self Cite

View full text Add to dashboard Cite

Many high-performance fluid dynamic models do not consider fluids in a rotating environment and often require a significant amount of computational time. The current study proposes a novel parameter-based field reconstruction convolutional neural network (PFR-CNN) approach to model the solute concentration field in rotationally influenced fluids. A new three-dimensional (3D) numerical solver, TwoLiquidMixingCoriolisFoam, was implemented within the framework of OpenFOAM to simulate effluents subjected to the influence of rotation. Subsequently, the developed numerical solver was employed to conduct numerical experiments to generate numerical data. A PFR-CNN was designed to predict the concentration fields of neutrally buoyant effluents in rotating water bodies based on the Froude number (Fr) and Rossby number (Ro). The proposed PFR-CNN was trained and validated with a train-validation dataset. The predicted concentration fields for two additional tests demonstrated the good performance of the proposed approach, and the algorithm performed better than traditional approaches. This study offers a new 3D numerical solver, and a novel PFR-CNN approach can predict solute transport subjected to the effects of rotation in few seconds, and the PFR-CNN can significantly reduce the computational costs. The study can significantly advance the ability to model flow and solute transport processes, and the proposed CNN-based approach can potentially be employed to predict the spatial distribution of any physical variable in the lentic, ocean, and earth system.

show abstract

Two-dimensional convolutional neural network outperforms other machine learning architectures for water depth surrogate modeling

Cited by 17 publications

References 43 publications

A dataset of storm surge reconstructions in the Western North Pacific using CNN

A dataset of storm surge reconstructions in the Western North Pacific using CNN

Surrogate-Assisted Evolutionary Algorithm for the Calibration of Distributed Hydrological Models Based on Two-Dimensional Shallow Water Equations

Fast Prediction of Solute Concentration Field in Rotationally Influenced Fluids Using a Parameter-Based Field Reconstruction Convolutional Neural Network

Contact Info

Product

Resources

About