Search results

In this paper, we present a comparison of linear and exponential interpolation functions for control volume finite element method. The exponential interpolation function is expressed in the elemental local coordinate system whereas the classic linear interpolation function is expressed in the global coordinate system. The comparison is achieved in the case of the Green‐Taylor vortex, a flow from which we know the analytical solution. Firstly, the two functions are applied to a triangular element of the domain to compare the results given by each interpolation function to the exact value. Secondly, these two functions are compared when used to solve the discretized equations over the entire domain.

A control‐volume based method for the numerical calculation of axisymmetric incompressible fluid flow and heat transfer is presented. The proposed method extends the staggered grid approach to unstructured triangular meshes. The velocities are stored at the vertices and the edges of a triangle, pressure and temperature are stored at the vertices. Accordingly, velocities are interpolated in a quadratic way, pressure and temperature linearly. The accuracy of the proposed method is examined for a number of different testproblems. Compared to a linear interpolation scheme implemented in the same code, more accurate solutions and smaller computation times are obtained for the proposed quadratic scheme. The method was designed for and is about to be applied to the numerical simulation of crystal growth.

The purpose of this paper is to develop a Vortex-In-Cell (VIC) method with the semi-Lagrangian scheme and apply it to the high-Re lid-driven cavity flow.

The VIC method is developed for simulating high Reynolds number incompressible flow. A semi-Lagrangian scheme is incorporated in the convection term to produce unconditional stability, which gets rid of the constraint of the convection Courant-Friedrichs-Lewy (CFL) condition; the adaptive time step is used to maintain the numerical stability of the diffusion term; and the velocity boundary condition is readily converted to the vorticity formulation to suit discontinuous boundary treatment. The VIC simulation results are compared with those produced by other gird methods reported in open literature studies.

The lid-driven cavity flow is simulated from Re = 100 to 100,000. Similar vortex birth mechanisms are exhibited though, but distinct flow characteristics are revealed. At Re = 100 to 7,500, the cavity flow is confirmed steady. At Re = 10,000, 15,000 and 20,000, the cavity flow is periodical with a primary vortex held spatially at the center. In particular, at Re = 100,000 highly turbulent characteristics is first revealed and an analogous primary vortex is formed but in motion rather than stationary, which is caused by the considerable flow separation at all the boundaries.

In the lid-driven cavity, the flow becomes extremely complex and highly turbulent at Re = 100,000, and the analogous primary vortex structure is observed. Boundary layer separation is observed at all walls, producing small vortices and causing the displacement of the analogous primary vortex. Such a finding original and has not yet been reported by other investigators. It may provide a basis for conducting in-depth studies of the lid-driven cavity flow.

A calculation procedure is proposed for heat and fluid flows in geometries with a time‐dependent boundary. The procedure incorporates a moving mesh arrangement with multi‐block iteration and has been developed to assist future simulations of heat and mass transfer with phase change. The solver for the basic equations is the SIMPLE algorithm with a non‐staggered grid arrangement. The space conservation law is invoked and applied for the explicit tracking of a moving boundary with a moving mesh. For mapping complex geometries a multi‐block iteration strategy is employed. A cubic spline interpolation allows the “uniqueness of zonal boundary” requirement to be met. An interpolation method is also developed for variables near the zone boundaries.The calculation procedure using multi‐block iteration and a moving mesh is applied to three benchmark‐testing problems. The numerical results are in very good general agreement with available experimental data.

This study aims to present compatible computational fluid dynamics procedure for calculation of incompressible three‐dimensional time‐dependent flow with complicated free surface deformation. A computer software is developed and validated using a variety of academic test cases.

Two fluids are modeled as a single continuum with a fluid property jump at the interface by solving a scalar transport equation for volume fraction. In conjunction, the conservation equations for mass and momentum are solved using fractional step method. Here, a finite volume discretisation and colocated arrangement are used.

The developed code results in accurate simulation of interfacial flows, e.g. Rayleigh‐Taylor instability, sloshing and dambreaking problems. All results are in good concordance with experimental data especially when there are two phases with high density ratio.

Turbulence, which has great importance in a wide variety of real world phenomena, is not considered in the present formulation and left for future researches.

Here, an integrated numerical simulation for transient interfacial flows is presented. In this way, the pressure integral term in Navier‐Stokes equation is discretised based on a newly developed interpolation which results in non‐oscillative velocity field especially in free surface.

Presents an interpolatory subdivision scheme to generate adaptively refined quadrilateral meshes which approximate a smooth surface of arbitrary topology. The described method differs significantly from classical mesh generation techniques based on spline surfaces or implicit representations since no explicit description of the limit surface is used. Instead, simple affine combinations are applied to compute new vertices if a face of the net is split. These rules are designed to guarantee asymptotic smoothness, i.e. the sequence of refined nets converges to a smooth limit surface. Subdivision techniques are useful mainly in applications where a given quadrilateral net is a coarse approximation of a surface and points on a refined grid have to be estimated. To evaluate the proposed approach, shows examples for FE‐computations on surfaces generated by this algorithm.

The purpose of this paper is to propose an effective and efficient numerical method that can consider natural frequency in multi-material topology optimization (MMTO) and which is scalable for complex three-dimensional (3D) problems.

The optimization algorithm is developed by combining custom FORTRAN code for MMTO with the open-source software Mystran, which is used as a finite element analysis (FEA) solver. The proposed algorithm allows the designer to shift the fundamental frequency of the design beyond a defined frequency spectrum from the initial designing phase. The methodology is formulated in a smooth and differentiable manner, with the sensitivity expressions, required by gradient-based optimization solvers, presented.

Natural frequency constraint has been successfully implemented into MMTO. The use of open-source software Mystran as an FEA solver in the algorithm provides ability to solve complex problems. Mystran offers powerful built-in functions for eigenvalue extraction using methods like Givens, modified Givens, inverse power and the Lanczos method, which provide the ability to solve complex models. The algorithm is successfully able to solve both two- and three-material MMTO jobs for two-dimensional and 3D geometries.

Natural frequency constraint consideration into topology optimization is very challenging due to three common issues: localized eigenmodes, mode switching and high computational cost. The proposed algorithm addresses these inherent issues, implements natural frequency constraint to MMTO and solves for complex models, which is hardly possible using conventional methods.

In 1982, Smith and Hutton published comparative results of several different convection‐diffusion schemes applied to a specially devised test problem involving near‐discontinuities and strong streamline curvature. First‐order methods showed significant artificial diffusion, whereas higher‐order methods gave less smearing but had a tendency to overshoot and oscillate. Perhaps because unphysical oscillations are more obvious than unphysical smearing, the intervening period has seen a rise in popularity of low‐order artificially diffusive schemes, especially in the numerical heat‐transfer industry. This paper presents an alternative strategy of using non‐artificially diffusive higher‐order methods, while maintaining strictly monotonic transitions through the use of simple flux‐limiter constraints. Limited third‐order upwinding is usually found to be the most cost‐effective basic convection scheme. Tighter resolution of discontinuities can be obtained at little additional cost by using automatic adaptive stencil expansion to higher order in local regions, as needed.

The present study aims to develop a vortex method capable for solving the complex vortical flows past the moving/deforming bodies.

To achieve such a goal, some innovative work is conducted on the basis of vortex-in-cell (VIC) method that uses the improved semi-Lagrangian scheme. The penalization technique is incorporated with the VIC, which makes the complex boundaries of moving/deforming bodies readily treated. Iterative algorithm is further proposed for the penalization and used to solve the Poisson equation, which enhances the vorticity solution accuracy at the body boundary.

The developed method is used to simulate some distinct flows of different boundaries and features: the impulsively started circular cylinder flow represents the one-way coupling; the falling circular cylinder flow and ellipse leaf flow both represent the two-way coupling of moving boundary; the fish-like body flow represents the two-way fluid-structure interaction of deforming boundary. The vortical physics of the above flows are well revealed, and the developed method is proven capable in dealing with the complex fluid-structure interaction problems.

The penalization technique is incorporated with the semi-Lagrangian VIC method, which makes the complex boundaries of moving/deforming bodies readily treated. An iterative algorithm is further proposed for the penalization and used to solve the Poisson equation, which enhances the vorticity solution accuracy at the body boundary. The complex vortical physics of the moving/deforming body flows are well revealed, and the propulsive mechanism of fish-like swimmer is well illustrated with the present method.

This paper aims to investigate the convergence and error properties of a finite volume-based heat conduction code that uses automatic differentiation to evaluate derivatives of solutions outputs with respect to arbitrary solution input(s). A problem involving conduction in a plane wall with convection at its surfaces is used as a test problem, as it has an analytical solution, and the error can be evaluated directly.

The finite volume method is used to discretize the transient heat diffusion equation with constant thermophysical properties. The discretized problem is then linearized, which results in two linear systems; one for the primary solution field and one for the secondary field, representing the derivative of the primary field with respect to the selected input(s). Derivatives required in the formation of the secondary linear system are obtained by automatic differentiation using an operator overloading and templating approach in C++.

The temporal and spatial discretization error for the derivative solution follows the same order of accuracy as the primary solution. Second-order accuracy of the spatial and temporal discretization schemes is confirmed for both primary and secondary problems using both orthogonal and non-orthogonal grids. However, it has been found that for non-orthogonal cases, there is a limit to the error reduction, which is concluded to be a result of errors in the Gauss-based gradient reconstruction method.

The convergence and error properties of derivative solutions obtained by forward mode automatic differentiation of finite volume-based codes have not been previously investigated.