Browsing by Author "Meyer, CJ"

Now showing 1 - 5 of 5
  • Thumbnail Image
    Item
    Open Access
    Development of a rotor model for the numerical simulation of helicopter exterior flow-fields
    (2004) Hotchkiss, Paul; Meyer, CJ; Von Backström, TW
    A numerical methodology is developed to model the effect of a rotor on the surrounding flow-field. The model calculates the time-averaged aerodynamic forces exerted on the air by the fan blades within the blade-swept region, and permits the user to specify blade properties such as cross-sectional profile and orientation at a particular radial and azimuthal location. The calculated forces are included as source terms within the Reynolds-averaged Navier-Stokes equations for an incompressible fluid, which are solved by the commercial CFD solver, FLUENT. The effects of turbulence are incorporated through the use of Launder and Spalding's k-g turbulence model. This method is selected as being the most efficient use of the resources available, giving the economic advantages of a steady simulation, while allowing radial and azimuthal variations of rotor characteristics. In order to validate the accuracy of the numerical model for both aligned and non-aligned inflow conditions, results are compared with experimental data reported for an axial flow fan. Agreement between experimental and numerical results is excellent to good. Fan static pressure rise is closely predicted by the numerical solution, while fan power consumption and fan static efficiency are under and over-predicted respectively. This error may be attributed to frictional losses not accounted for in the numerical model. These include physical rotational instabilities, leading to increased mechanical losses, and tip effects due to the clearance between the fan blade tips and the fan casing. Trends are nevertheless consistently predicted by the numerical model for inflow angles up to 45°, and for the range of blade pitch settings used. The adverse effect of off-axis inflow on the fan static pressure rise is numerically predicted, while fan power consumption is found to remain independent of inflow angle, as had been experimentally observed. The rotor model is finally integrated with the fuselage of the CIRSTEL (Combined Infra-Red Suppression and Tail rotor Elimination) prototype in an analysis of the helicopter exterior flow-field. No experimental data for this configuration was available for validation purposes. However, the model is used in the simulation of several common helicopter flight conditions. Results are presented graphically, and generally indicate good agreement with physically observed phenomena.
  • Thumbnail Image
    Item
    Open Access
    Implementation of a trim routine in a rotor model for the numerical simulation of helicopter flow-fields
    (2006) Naidoo, Vaneshen; Meyer, CJ
    The aim of the current project is to develop, validate and implement a trim routine for a numerical rotor model, developed for the use in simulations of a helicopter exterior flow-field. In this investigation a ROBIN fuselage geometry was utilised. Simulations of the fuselage without the rotor were carried out initially so that investigations into the computational grids and turbulence models could be done. The computational simulations were performed in the commercially available CFD solver, FLUENT® Computational grids were created for the near wall modelling approach and wall function approach. Some of the more applicable turbulence models available in the solver were compared. For the wall function approach grids the k - ε, and its variants, the RNG and realizable models were found to be suitable choices. For the near wall modelling approach grids used, the SST models performed the best. The rotor model used during this investigation utilised a combination of blade element and actuator disk theory. Forces exerted by the rotor are calculated with the use of blade characteristics and flow properties. These forces were applied to the domain as momentum sources terms. The rotor model was incorporated with the CFD solver, through the use of a User Defined Function (UDF). The method used to trim the rotor was the Newton-Raphson Iterative method. This trim routine was incorporated in the UDF used for the rotor model. Tests were conducted, on a 'rotor-alone' model, as well as the rotor and fuselage model. The trim routine was found to be rigorous and managed to trim the rotor in each of the tests conducted. Good agreement between experimental and numerical collective pitch angle and cyclic pitch coefficients were found. Also the effect of the fuselage on the trim conditions proved to be minimal.
  • Thumbnail Image
    Item
    Open Access
    Multiphase CFD modelling of stirred tanks
    (2007) Appa, Harish; Deglon, David; Meyer, CJ
    Stirred tanks agitated with Rushton turbines are commonly used in industry, for instance mixing processes and flotation systems. The need for more efficient systems in industries has led to the study of fluid flow within the tanks upon agitation; so that a better understanding of the phenomena can help in the optimisation of the tanks. In the recent years, efforts have been made towards the development of predictive methods using computational fluid dynamics (CFD). Among the various numerical works presented, emphasis was laid mainly on single phase systems. However, due to the various processes involving gas-liquid systems, the need for multiphase modelling of stirred tanks became increasingly important. This has led to more research studies involving multiphase flows. Most of the work reported showed good prediction of the velocity data and the power draw, reasonable turbulence parameters. But, the prediction of the gas hold-up was rarely well established. Therefore, the aim of this thesis, based on the numerical work presented by Engelbrecht (2006), is to investigate the discrepancies reported and to develop a multiphase model of a stirred tank agitated by a Rushton turbine. The commercially available CFD code FLUENT@ was used to model the agitated gas-liquid system. The results were validated with the numerical work of Engelbrecht (2006) and the experimental work presented by Deglon (1998). Two main cases were investigated, with a steady state and a transient approach. The QUICK scheme was used for the discretisation of the volume fraction and momentum and the first order upwind scheme for the discretisation of the turbulent kinetic energy and dissipation rate. The standard k - E turbulence model was used to account for the turbulent flow regime. A steady state MRF model was used for the investigation of the discrepancy reported by Engelbrecht (2006). The author reported that no convergence was achieved with such models. Solving the problem would have resulted in a good modelling approach for the prediction of gas dispersion, since steady state models are not computationally intensive. Three different boundary conditions, namely, a pressure outlet, an outflow and a velocity inlet, were used to model the outlet of the tank. The Euler-Euler multiphase model was used to simulate the gas-liquid system for the steady state model.
  • Thumbnail Image
    Item
    Open Access
    Numerical modelling of hydrodynamics, gas dispersion and mass transfer in an autoclave
    (2012) Appa, Harish; Deglon, David; Meyer, CJ
    Includes abstract. Includes bibliographical references.
  • Thumbnail Image
    Item
    Open Access
    Simulation of fluid suspended particle behaviour subject to transverse standing acoustic fields
    (2012) Dabic, Mihajlo; Deglon, David Alan; Meyer, CJ
    The computational study addressed the effectiveness with which a standing wave acoustic field could be used to deflect quartz particles carried in water at 20°C through a simple parallelepiped control volume representative of a vertically orientated duct geometry dimensioned 50 × 50 × 70 cm 3 , with square base. An acoustically driven planar standing wave field produces quasi-static oscillatory pressure gradients within resonant cavities, which are responsible for acoustic forces, which act on particles, located within the acoustic field. These forces drive particles to nodal (no fluctuation) or anti-nodal (continuous fluctuation) planes of pressure. Standing wave fields are generally produced by a transducer driving into a fluid through an adhesively bonded matching layer. The wave is reflected at the opposite boundary layer terminating in an air backing. The chamber is dimensioned so as to produce constructive wave interference between the two waves travelling in opposite directions. The acoustic force has been used in small scale filtration systems to deflect particles and on larger scales as a pre filtration agglomerator clumping very small particles which are otherwise poorly filtered in isolation by conventional methods. The study was twofold, in that a major component of the study comprised developing the architecture of the computational model, the other part comprising qualitative model validation through parameter variation. The study involved coupling between Computational Fluid Dynamics (CFD) Software (OpenFOAM) and Discrete Element Modelling (DEM) Software (LIGGGHTS), through a coupling code (CFDEM) built as an extension to OpenFOAM and tailored for LIGGGHTS. The acoustic field was assumed ideal i.e. in a lossless medium with perfect reflection at the opposite wall. Particle-particle and particle-wall collisions were circumvented by using larger time increments, inadequate to resolve col- lisions, and inserting particles in the bulk of the flow away from any wall boundary. Twenty particles with uniform radial size distribution in the range 5-30 micron were seeded in the flow field about 10 cm from the bottom inlet, and carried in the z direction at various flow speeds, 0.1 ms − 1 , 0.5 ms − 1 and 1 ms − 1 , whilst being subject to acoustic forces in the x direction, to investigate deflection response and transducer lengths required to achieve adequate lateral deflection. The model accounted for drag, buoyancy, gravity and primary acoustic forces. Flow velocities distinguished by those maxi mum velocities recorded at duct centrelines were obtained by adjusting pressure gradients across the domain. The fluid continuum was modelled through Reynolds Averaged Navier Stokes (RANS) equa- tions, supplemented by an eddy viscosity k − two equation turbulence model. The flow profile was validated against the analytic Darcy-Weisbach pressure to mean velocity relation. Two acoustic driv- ing frequencies, 14794 Hz and 26629 Hz , were investigated for each flow rate to determine the effect frequency had on acoustic force magnitude, nodal distribution and particle residence time. Acoustic deflection efficiency was measured as that time or particle vertical travel length required, coinciding with a lateral deflection to within 1.5 mm of an adjacent nodal plane. From a computational point of interest acoustic force dependencies and trends were qualitatively evaluated for consistency with theoretic equations and published literature