Zexia Zhang

Flood‐induced deformation of the bed topography of fluvial meandering rivers could lead to riverbank displacement, infrastructure failure, and the propagation of scour and deposition features. The ability to predict sediment transport in large‐scale meanders is, therefore, a critical requirement for tackling a host of environmental issues. High‐fidelity simulations of large‐scale rivers coupling large‐eddy simulation (LES) of the flow with morphodynamic processes can provide accurate predictions of such phenomena, but this approach is computationally very expensive owing to the costly two‐way coupling between turbulence and bed morphodynamics across disparate time scales. This study presents a novel machine learning approach trained on data from high‐fidelity simulations. We develop and demonstrate the performance of a convolutional neural network autoencoder (CNNAE) algorithm to generate high‐fidelity bed shear stress and equilibrium morphology of large‐scale meandering rivers. The CNNAE algorithm utilizes as input instantaneous shear stress distribution and change of bed elevation obtained from high‐fidelity simulation results, along with geometric parameters of meanders to predict mean bed shear stress distribution and equilibrium bed elevation of rivers. The proposed machine learning approach predicts bed shear stress and equilibrium bed morphology of large‐scale meanders under bankfull flow conditions at a fraction of the cost of brute‐force coupled LES‐morphodynamics.

This study proposes a novel machine learning (ML) methodology for the efficient and cost-effective prediction of high-fidelity three-dimensional velocity fields in the wake of utility-scale turbines. The model consists of an autoencoder convolutional neural network with U-Net skipped connections, fine-tuned using high-fidelity data from large-eddy simulations (LES). The trained model takes the low-fidelity velocity field cost-effectively generated from the analytical engineering wake model as input and produces the high-fidelity velocity fields. The accuracy of the proposed ML model is demonstrated in a utility-scale wind farm for which datasets of wake flow fields were previously generated using LES under various wind speeds, wind directions, and yaw angles. Comparing the ML model results with those of LES, the ML model was shown to reduce the error in the prediction from 20% obtained from the Gauss Curl hybrid (GCH) model to less than 5%. In addition, the ML model captured the non-symmetric wake deflection observed for opposing yaw angles for wake steering cases, demonstrating a greater accuracy than the GCH model. The computational cost of the ML model is on par with that of the analytical wake model while generating numerical outcomes nearly as accurate as those of the high-fidelity LES.

We present a physics-enhanced convolutional neural network (PECNN) algorithm for reconstructing the mean flow and turbulence statistics in the wake of marine hydrokinetic (MHK) turbine arrays installed in large-scale meandering rivers. The algorithm embeds the mass and momentum conservation equations into the loss function of the PECNN algorithm to improve the physical realism of the reconstructed flow fields. The PECNN is trained using large eddy simulation (LES) results of the wake flow of a single row of turbines in a virtual meandering river. Subsequently, the trained PECNN is applied to predict the wake flow of MHK turbines with arrangements and positionings different than those considered during the training process. The PECNN predictions are validated using the results of separately performed LES. The results show that the PECNN algorithm can accurately predict the wake flow of MHK turbine farms at a small fraction of the cost of LES. The PECNN can improve the accuracy by around 1% and reduce the physical constraint indices by around 50% compared to the CNN without physical constraints. This work underscores the potential of PECNN to develop reduced-order models for control co-design and optimization of MHK turbine arrays in natural riverine environments.

In meandering rivers, interactions between flow, sediment transport, and bed topography affect diverse processes, including bedform development and channel migration. Predicting how these interactions affect the spatial patterns and magnitudes of bed deformation in meandering rivers is essential for various river engineering and geoscience problems. Computational fluid dynamics simulations can predict river morphodynamics at fine temporal and spatial scales but have traditionally been challenged by the large scale of natural rivers. We conducted coupled large‐eddy simulation and bed morphodynamics simulations to create a unique database of hydro‐morphodynamic data sets for 42 meandering rivers with a variety of planform shapes and large‐scale geometrical features that mimic natural meanders. For each simulated river, the database includes (a) bed morphology, (b) three‐dimensional mean velocity field, and (c) bed shear stress distribution under bankfull flow conditions. The calculated morphodynamics results at dynamic equilibrium revealed the formation of scour and deposition patterns near the outer and inner banks, respectively, while the location of point bars and scour regions around the apexes of the meander bends is found to vary as a function of the radius of curvature of the bends to the width ratio. A new mechanism is proposed that explains this seemingly paradoxical finding. The high‐fidelity simulation results generated in this work provide researchers and scientists with a rich numerical database for morphodynamics and bed shear stress distributions in large‐scale meandering rivers to enable systematic investigation of the underlying phenomena and support a range of river engineering applications.

Prediction of statistical properties of the turbulent flow in large‐scale rivers is essential for river flow analysis. The large‐eddy simulation (LES) provides a powerful tool for such predictions; however, it requires a very long sampling time and demands significant computing power to calculate the turbulence statistics of riverine flows. In this study, we developed encoder‐decoder convolutional neural networks (CNNs) to predict the first‐ and second‐order turbulence statistics of the turbulent flow of large‐scale meandering rivers using instantaneous LES results. We train the CNNs using a data set obtained from LES of the flood flow in a large‐scale river with three bridge piers—a training testbed. Subsequently, we employed the trained CNNs to predict the turbulence statistics of the flood flow in two different meandering rivers and bridge pier arrangements—validation testbed rivers. The CNN predictions for the validation testbed river flow were compared with the simulation results of a separately done LES to evaluate the performance of the developed CNNs. We show that the trained CNNs can successfully produce turbulence statistics of the flood flow in the large‐scale rivers, that is, the validation testbeds.

A convolutional neural network (CNN) autoencoder model has been developed to generate 3D realizations of time-averaged velocity in the wake of the wind turbines at the Sandia National Laboratories Scaled Wind Farm Technology (SWiFT) facility. Large-eddy simulations (LES) of the SWiFT site are conducted using an actuator surface model to simulate the turbine structures to produce training and validation datasets of the CNN. The simulations are validated using the SpinnerLidar measurements of turbine wakes at the SWiFT site and the instantaneous and time-averaged velocity fields from the training LES are used to train the CNN. The trained CNN is then applied to predict 3D realizations of time-averaged velocity in the wake of the SWiFT turbines under flow conditions different than those for which the CNN was trained. LES results for the validation cases are used to evaluate the performance of the CNN predictions. Comparing the validation LES results and CNN predictions, we show that the developed CNN autoencoder model holds great potential for predicting time-averaged flow fields and the power production of wind turbines while being several orders of magnitude computationally more efficient than LES.