Various regions are becoming increasingly vulnerable to the increased frequency of floods due to the recent changes in climate and precipitation patterns throughout the world. As a result, specific infrastructures, notably bridges, would experience significant flooding for which they were not intended and would be submerged. The flow field and shear stress distribution around tandem bridge piers under pressurized flow conditions for various bridge deck widths are examined using a series of three-dimensional (3D) simulations. It is indicated that scenarios with a deck width to pier diameter (Ld/p) ratio of 3 experience the highest levels of turbulent disturbance. In addition, maximum velocity and shear stresses occur in cases with Ld/p equal to 6. Results indicate that increasing the number of piers from 1 to 2 and 3 results in the increase of bed shear stress by 24% and 20% respectively. Finally, five machine learning algorithms, including Decision Trees (DT), Feed Forward Neural Networks (FFNN), and three Ensemble models, are implemented to estimate the flow field and the turbulent structure. Results indicated that the highest accuracy for estimation of U, and W, were obtained using AdaBoost ensemble with R² = 0.946 and 0.951, respectively. Besides, the Random Forest algorithm outperformed AdaBoost slightly in the estimation of V and turbulent kinetic energy (TKE) with R² = 0.894 and 0.951, respectively.

Conventional digital twin platforms typically rely on the sensors to collect monitoring data at single points, thus making it difficult to capture the complex flow characteristics. In this study, a five-dimensional digital twin platform was established for the lateral inlet/outlet of a pumped storage power station. Numerical simulation was integrated to obtain the spatiotemporal distribution of the entire flow field. The platform was composed of five layers—physical entity, digital twin, connection, data and service layer. Intelligent fluid–structure simulation analysis was combined with the real-time rendering of complex data. The velocity distribution and trash rack modal patterns were within an immersive virtual reality environment. Real-time flow field analysis of the inlet/outlet was provided for the online monitoring of one-dimensional basic data, such as the velocity and discharge at each orifice, as well as complex two- and three-dimensional data, including cross-sectional velocity distribution and flow pattern evolution. High-precision flow field results show that the main flow was separated from the wall surface at both horizontal and vertical diffusion angles. A recirculation zone was formed at the top of the diffuser section, where the main stream was concentrated in the lower part of the channel. There was very low turbulence intensity in the main flow and recirculation regions. While two shear layers with relatively high turbulence intensity were observed above and below the main flow region in the streamwise direction from the beginning of the diffuser section to the trash rack cross-section. These shear layers were corresponded to small-scale vortex motions, indicating a strong correlation between vortex distribution and turbulence intensity. Modal analysis of the trash rack structure revealed that the first and second modes were mainly characterized by overall vibration of the rack frame along the z-direction, resulting from the bending modes of the main and secondary beams. In the first mode, the bars near the side beams also exhibited the outstanding torsion. While in the second mode, the torsion occurred throughout the entire bar system. The fifth mode often presented a coupled bending deformation with the bars, side beams, and longitudinal beams, together with the bar torsion. This complex vibration pattern was easily led to the large deformation and potential structural failure of the bar system. However, the main and secondary beams were supported by the bars and longitudinal beams. The high stiffness was less prone to damage. Moreover, in the lower-order modes, the optimal thickness parameters were effectively prevented to damage, where the bars shared no bending patterns that posed a serious threat to structural safety. The frequency ratio between the fifth-order mode and the vortex-shedding frequency was ranged from 2.38 to 4.58, with the lower bound falling below the code-specified threshold of 2.5. Therefore, the special attention should be given to the potential for the flow-induced excitation of the fifth-order mode of the trash rack in engineering practice, leading to the resonance and structural damage near the wall surface. The platform can be expected to realize the high efficiency and rapid feedback in the hydraulic optimization using intelligent numerical simulation of lateral inlet/outlet structures and trash racks.