Abstract:This study presents an artificial neural network and proper orthogonal decomposition (POD)-based reduced-order model (ROM) of turbulent flow around a finite wall-mounted square cylinder. The proposed model is suitable for turbulent wake control applications because it can predict the dynamics of the main features of the flow field without computing Navier-Stokes equations. Long short-term memory neural network (LSTM NN) and bidirectional long short-term memory neural network (BLSTM NN) are used to predict the temporal evolution of the POD time coefficients at different planes along the height of the obstacle. The improved delayed detached-eddy simulation (IDDES) is performed to generate the training datasets. Transfer learning (TL) approach is utilized in the training process by using the weights of the LSTM/BLSTM NN that are used to predict the POD time coefficients of the planes at lower elevations to initialize the weights of the networks at higher elevations along the height of the obstacle. The use of TL results in a remarkable improvement in the capability of the LSTM/BLSTM NN prediction compared with the one when the network is initialized with random weights. BLSTM NN shows better results compared with LSTM NN in terms of training and prediction error, indicating that the BLSTM-POD model is more suitable to be used as a ROM for predicting the turbulent wake. Furthermore, the temporal behavior of the time coefficients is carefully examined using the phase space plots and Poincar$\acute{e}$ sections. The results of using different lengths of the prediction time window showed that the prediction error of the POD time coefficients increases as the prediction time window increases and the error increasing rate decreases with the ranking of the POD time coefficients.

Three-dimensional wake transition of rectangular cylinders and temporal prediction of flow patterns based on a machine learning algorithm

Three-dimensional wake transitions of steady flow past two side-by-side cylinders

Three-dimensional Direct Numerical Simulation of Wake Transitions of a Circular Cylinder

A Hybrid Reduced-Order Model Combing Deep Learning for Unsteady Flow

Three-dimensional Numerical Simulation on Flow Past Three Circular Cylinders in an Equilateral-Triangular Arrangement

Numerical Simulation of Three-Dimensional Vortex Dynamics in Wake of A Circular Cylinder

Numerical Exploration of New Features on Three-Dimensional Transition of a Cylinder Wake

Artificial neural network-based reduced-order modeling for turbulent wake of a finite wall-mounted square cylinder

Two- and Three-Dimensional Wake Transitions of a Rectangular Cylinder and Resultant Hydrodynamic Effects

Three-Dimensional Flow Around Two Circular Cylinders in Tandem Arrangement

Three-Dimensional Flow Around Two Tandem Circular Cylinders with Various Spacing at Re=220

Unsteady Cylinder Wakes from Arbitrary Bodies with Differentiable Physics-Assisted Neural Network

Three-Dimensional Wake Transition for A Circular Cylinder Near A Moving Wall

Mode interpretation and force prediction surrogate model of flow past twin cylinders via machine learning integrated with high-order dynamic mode decomposition

The near wake of sinusoidal wavy cylinders: Three-dimensional POD analyses

A Comparative Study of Data-Driven Modal Decomposition Analysis of Unforced and Forced Cylinder Wakes

Wake of two tandem square cylinders

Three-dimensional Wake Transition in the Flow over Four Square Cylinders at Low Reynolds Numbers

Transitions in the Unsteady Wakes and Aerodynamic Characteristics of the Flow Past Three Square Cylinders Aligned Inline

Bidirectional prediction between wake velocity and surface pressure using deep learning techniques