For clarification of an ecological system or application to ship propulsion mechanisms, studies of fish swimming have been conducted. However, many parts remain unexplained due to complexity and diversity of the swimming mechanism. In this study, we focused on the shape change of the fish larvae’s fin fold and investigated effects of aspect ratio of fish larvae’s fin fold on its propulsion performance by numerical simulations. Flow around the fish model was simulated by the regularized lattice Boltzmann method. To describe curved boundary on a Cartesian grid, the virtual flux method was applied. 4-tired multi block method was employed, and calculation domain was divided into 4 blocks with different grid resolutions. In this simulation, aspect ratio was defined by the ratio of fin fold length to fin fold width at the tail end, and fish models with different aspect ratio of fin fold were applied under constant surface area condition. As a result, vortices were generated along the fish body, and edge vortices pairs rotating in the opposite direction were observed at the fin fold. The edge vortices contributed to generate thrust force by increasing the front-back pressure difference at the fin fold. Compared to lower aspect ratio models, higher aspect ratio models generated more edge vortices pairs during a cycle. Accordingly, at lower aspect ratio, thrust force became larger by the wider fin fold, but amplitude of swimming speed changes in a cycle also became larger. In contrast, higher aspect ratio models swam efficiently by stabilizing its swimming speed.
We carried out the implementation of visualization techniques that aim at the interpretation of the complex turbulent structures in fluid flow. Turbulent structures help in understanding various phenomena of the flow field such as energy dissipation and drag reduction. However, few studies were found regarding vortex sheet structure identification for experimental data. In the current study, these sheet structures were visualized. Experimental data for the three-dimensional instantaneous velocity vector field was measured using the Tomographic PIV method. Two different kinds of flow settings, pipe flow and whirlpool flow, were measured and analyzed. For comparison of results, pipe measurements were recorded at varying Reynolds number and whirlpool flow were recorded with constant stirring and decaying turbulence. Q criterion analysis, which evaluates regions of higher vorticity than strain, was employed for visualization of vortex tubes in a time series flow field measurement. Vortex sheets were visualized using the Aij matrix analysis that utilizes second-order velocity gradient tensor and its eigenvalues for identifying regions with a high correlation and magnitudes of vorticity and strain. The relative position of the tube and sheet structures visualized using the two methods were reviewed for the experimental data of pipe flow and whirlpool flow. Comparison of the experimentally resolved structures with structures of DNS data was studied for pipe flow and a close resemblance was observed.
This paper presents a shape optimazation method for transient Non-Newtonian fluid which is playing important roles of calculating blood flow, oil flow and so on. So far, the author constructed a shape optimization problem for suppressing transient Newtonian fluid by using Snapshot POD, and extends it toward to Non-Newtonian fluid, here. For such the suggested shape optimization, the eigenvalue in Snapshot POD is defined as a cost function, where the constraint functions are the Oldroyd-B model, and an eigenvalue equation of Snapshot POD. For numerical calculations, a two-dimensional cavity flow with a disk-shaped isolated body is adopted for an initial domain. To descritize the Oldroyd-B model spatially, Galerkin Method (GM) and Hybridized Discontinuous Galerkin Method (HDGM) are used to compare numerical accuracies. As a result, it is considered that HDGM is able to obtain better solutions than GM during numerical validations. Finally, eigenvalues of Snapshot POD are compared in the initial and optimal domains obtained by HDGM.
Einstein’s viscosity formula is sometimes strongly limited for viscosity estimation of suspensions; that is, it is only applicable for low-concentration suspensions in which hydrodynamic interactions are sufficiently negligible. In particular, hydrodynamic interactions between particles (cylinders in two dimensions) should be taken into consideration when finite-size particles are suspended. Therefore, change in the microstructure, i.e., spatial arrangement of particles in the flow field, is important for understanding mechanism of suspension rheology. In order to provide better practical applications for viscosity estimation instead of Einstein’s formula, we investigated the influence of each cylinder’s contribution on the total effective viscosity of a suspension with finite-size cylinders considering the microstructure, especially in terms of cylinder-wall and cylinder-cylinder distances. Two-dimensional pressure-driven flow simulations were performed using the regularized lattice Boltzmann method and a two-way coupling scheme. The rigid circular cylinders suspended in a Newtonian fluid were assumed to be neutrally buoyant and non-Brownian. As a result, we found that both distances between cylinders and cylinder-wall are significant for viscosity estimation. In addition, the effective viscosity can be estimated accurately when the confinement is sufficiently low (C ≈ 0.04). It can be stated that the microstructure of the suspension is one of the promising factors to estimate and control suspension rheology.
Three-dimensional density measurement of unsteady flow field around a projectile (φ = 8mm) is carried out in the ballistic range at Institute of Fluid Science, Tohoku University. The purpose of this study is to obtain the density distribution including its wake region in non-axisymmetric unsteady flow. The projectile’s Mach number in the experiment is 1.35. Simultaneous multi-angle BOS measurement system using twelve digital cameras and pulsed LEDs is installed in the test chamber of the ballistic range. The Color-Grid Background Oriented Schlieren (CGBOS) technique is used in the measurement system to obtain the projection data of density gradient. Algebraic Reconstruction Technique (ART) method and Successive Over Relaxation (SOR) method are used to obtain the density distribution from projection data. To improve the accuracy of 3D reconstruction, we evaluate the view-angle of the measurement system in the present study. The result shows that the bow shock, expansion wave and recirculation zone can be confirmed by the proposed method. However, bow shock in each method are thicker than the schlieren image due to the blur of the CGBOS images. Even then, the 3D shape of density distribution reconstructed by proposed method can be reliable, and the relative error of obtained density between the theoretical value for normal shock relation is about 5% at the bow shock.
It is important to understand the rheology of suspensions because it has a wide range of relevance from biological to industrial fields. The rheology of suspensions is still unclear due to complexity of various factors. Among the factors that determine the rheological properties, we focused on the spatial arrangement of particles and solvent properties of the suspension. We investigated effects of the power-law fluidic properties of the solvent of suspension on the relative and intrinsic viscosities. Furthermore, we investigated the effect of Reynolds number on the rheology of the suspension. We performed a numerical simulation of pressure-driven suspension flows in 2D. The suspension had different solvents properties depends on the power-law model. The bulk flow was simulated by using the lattice Boltzmann method. The power-law model was used to represent the flow properties of the solvent. The particle shape was described on the Cartesian grid using the virtual flux method. The relative and intrinsic viscosities of the suspensions were discussed by property changes in the suspension rheology such as shear-thinning, Newtonian, and shear-thickening. The results showed that the higher power-law index of the solvent caused higher relative and intrinsic viscosities. Furthermore, Reynolds number had little influence on the relative and intrinsic viscosities of the suspension when the Reynolds number was under Re = 12.