The hydrodynamic interaction between a main vehicle (MV) and an unmanned underwater vehicle (UUV) during their side-by-side navigation affects the stern flow of the main vehicle, thereby altering the inflow conditions at the propeller plane. To improve the flow uniformity at the propeller plane of the main vehicle, jet holes on the wall of the main vehicle are proposed to actively control the flow near the stern. Numerical simulations are conducted to investigate the influence of the jet on the flow at the propeller plane. The results show that the jet angle affects the direction of momentum transport, and a high-relative-intensity jet enhances the momentum transport process. The flow structures in the low-velocity zone near the blade root and the velocity magnitude at the blade tip are jointly controlled by the jet angle and relative intensity. This tuned jet enables active control of flow uniformity at the propeller plane. For the flow around the MV-UUV, a favorable match of jet angle and relative intensity has been identified to enhance flow uniformity at the propeller plane by approximately 19
Pump cavitation can lead to a significant decline in the pump’s performance and material erosion. This study aims to investigate the influence mechanism of non-uniform inlet flow on the cavitation behavior within a hydrodynamically-levitated micropump, utilizing computational fluid dynamics (CFD) simulations to analyze the cavitation flow dynamics and conducting experimental tests for validation. To facilitate a more comprehensive comparison of the cavitation performance of suspended pumps under varying inlet conditions, three distinct suspended pumps with different inlet configurations were compared and analyzed. The results indicate that the low-pressure region within the suspended pump initially forms near the leading edge of the vane. Cavitation-induced bubbles predominantly accumulate in the regions near the leading edge of the vertical section of the volute, the impeller inlet plane, the leading edge of the impeller’s suction side, and the root of the impeller. Specifically, cavitation at the leading edge of the vertical section of the volute and the impeller inlet plane is primarily influenced by vortex formation resulting from the change in the direction of the incoming flow and the inherent fluid inhomogeneity. Increasing the diameter of the volute inlet and employing a straight pipe inlet at the head of the suspended pump have minimal impact on the hydraulic performance. However, these modifications significantly enhance anti-cavitation performance, primarily by reducing vortex formation and fluid inhomogeneity. Experimental results demonstrate that increasing the diameter of the volute inlet leads to a 5.6
Submerged structures significantly influence sediment transport in open channels. In this study, the immersed boundary method (IBM) was integrated into a three-dimensional hydrodynamic simulation to enhance the numerical model of bedload transport in the presence of submerged obstacles. The feasibility of the improved model was validated using a submerged-cylinder flume experiment and a scaled model of a lateral intake on the Rong River, China. The improved model showed better representation of the submerged obstacles and thus more accurate simulation of the surrounding flow field, including hydrodynamic properties and sediment erosion processes. Moreover, the IBM-enabled method overcame the mesh generation limitations to save computational resources. Compared to conventional numerical models, the enhanced numerical model improved accuracy by at least 50
High-performance computing is critical in hydrodynamic simulations of floods, where the grid resolution and floating-point format significantly affect computational efficiency. This study employs a self-developed model to simulate hydrodynamic processes, using multi-graphics processing unit (GPU) parallel technology to enhance the computation speed. Within the multi-GPU architecture, the computational domain is partitioned using a ray-tracing method, and a data communication mechanism is designed to facilitate data transfer among GPUs. Different grid partitioning schemes and floating-point formats (FP64, FP32, FP16) are applied to analyze the model’s computational efficiency and influencing factors. The results indicate that multi-GPU parallel computing significantly enhances computational efficiency. Under a dual-GPU configuration, the model achieves up to 465× speedup compared with a CPU serial model, with a peak speedup of 1.79× relative to a single-GPU parallel model. The grid resolution greatly impacts the parallel computing efficiency. As the grid resolution increases, the computational load grows substantially, which contributes to promoting the advantages of multi-GPU parallel processing. However, the communication overhead rises correspondingly, which tends to limit the speedup. The choice of floating-point format requires balancing error control and computational efficiency. FP32 ensures sufficient accuracy and achieves a significant improvement in computational efficiency compared to FP64, demonstrating clear advantages in flood simulations. The proposed model based on multi-GPU parallel computing effectively uses hardware resources to support shallow water modelling.
A pump as turbine (PAT) is a cost-effective energy recovery device widely utilized in micro-hydropower and industrial systems. However, its hydraulic performance is significantly impacted by unsteady flow phenomena such as flow separation, vortices, and recirculation. Based on the spatial topological structures and statistical characteristics derived from steady-state simulations, this study conducts a quantitative correlation analysis between the Liutex method and entropy generation theory, revealing for the first time the physical mechanism of turbulence energy dissipation dominated by shear layers within this system. It also elucidates the complex role of vortex structures in modulating flow, as these structures act as more than just direct contributors to energy loss. The results demonstrate a distinct spatial correspondence between vortex structure distribution and entropy generation characteristics. By employing the “Vortex Strength-Entropy Generation” phase diagram, a three-stage energy dissipation cascade within the turbine is quantitatively revealed for the first time. This cascade commences with shear-dominated losses at the volute inlet, proceeds through a turbulence “singularity” during the impeller’s energy extraction process, and culminates in a dissipative endpoint marked by spatial development of a vortex in the suction chamber. This research highlights the advantages of the Liutex method in vortex identification and provides a new theoretical basis for future targeted drag reduction and efficiency enhancement designs.
