No description available.

No description available.

Commercial pressures to reduce time-scales encourage innovation in the design and analysis cycle of gas turbine combustion systems. The migration of Computational Fluid Dynamics (CFD) from the purview of the specialist into a routine analysis tool is crucial to achieve these reductions and forms the focus of this research. Two significant challenges were identified: reducing the time-scale for creating and solving a CFD prediction and reducing the level of expertise required to perform a prediction. The commercial pressure for the rapid production of CFD predictions, coupled with the desire to reduce the risk associated with adopting a new technology led, following a review of available techniques, to the identification of structured grids as the current optimum methodology. It was decided that the task of geometry definition would be entirely performed within commercial Computer Aided Design (CAD) systems. A critical success factor for this research was the adoption of solid models for the geometry representation. Solids ensure consistency, and accuracy, whilst eliminating the need for the designer to undertake difficult, and time consuming, geometry repair operations. The versatility of parametric CAD systems were investigated on the complex geometry of a combustion system and found to be useful in reducing the overhead in altering the geometry for a CFD prediction. Accurate and robust transfer between CAD and CFD systems was achieved by the use of direct translators. Restricting the geometry definition to solid models allowed a novel two stage grid generator to be developed. In stage one an initial algebraic grid is created. This reduces user interaction to a minimum, by the employment of a series of logical rules based on the solid model to fill in any missing grid boundary condition data. In stage two the quality of the grid is improved by redistributing nodes using elliptical partial differential equations. A unique approach of improving grid quality by simultaneously smoothing both internal and surface grids was implemented. The smoothing operation was responsible for quality, and therefore reduced the level of grid generation expertise required. The successful validation of this research was demonstrated using several test cases including a CFD prediction of a complete combustion system.

No description available.

The prediction of shut-off head within a centrifugal pump or fan has historically relied on the use of a combination of empirical formulae, together with experience. This reliance has been forced on designers due to both a lack of information regarding the flow at shut-off and the available tools with which to gain more accurate insights. To improve understanding the following investigation examines the capability of the commercial CFD package, FLUENT, to model the flow in a centrifugal pump at shut-off conditions. The computational model was validated using experimental measurements from a purpose built two-dimensional centrifugal pump rig. The rig used air as the working fluid and was similar in design to that used by Miner. Measurements were made of both the fluctuating velocities, using LDA, and the fluctuating pressures, using microphones, within the volute of the pump. The CFD model uses a sliding mesh which enables the full time-dependent rotor/stator interaction of the pump to be modelled. The results show the volute flow contains two patterns, a recirculating eddy in the outlet duct and a volute flow circulating around the rotor. This volute flow separates partway around the volute, with re-attachment on the discharge side of the tongue. The major effect of the volute at shut-off is to act as a diffuser with a strong circumferential pressure gradient over approximately the first 1200 after the tongue. A comparison of the experimental and computational results showed that good qualitative agreement was obtained at most positions at shut-off and considerable insight was gained into the flow mechanisms. However, the results showed that the CFD model over predicted the measured shut-off head by 25%. This was attributed to an over prediction of the effective viscosity due to the use of the k-s turbulence model.

This thesis presents new tools and techniques developed to address the challenging problem of high-fidelity aerostructural optimization with respect to large numbers of design variables. A new mesh-movement scheme is developed that is both computationally efficient and sufficiently robust to accommodate large geometric design changes and aerostructural deformations. A fully coupled Newton-Krylov method is presented that accelerates the convergence of aerostructural systems and provides a 20% performance improvement over the traditional nonlinear block Gauss-Seidel approach and can handle more flexible structures. A coupled adjoint method is used that efficiently computes derivatives for a gradient-based optimization algorithm. The implementation uses only machine accurate derivative techniques and is verified to yield fully consistent derivatives by comparing against the complex step method. The fully-coupled large-scale coupled adjoint solution method is shown to have 30% better performance than the segregated approach. The parallel scalability of the coupled adjoint technique is demonstrated on an Euler Computational Fluid Dynamics (CFD) model with more than 80 million state variables coupled to a detailed structural finite-element model of the wing with more than 1 million degrees of freedom. Multi-point high-fidelity aerostructural optimizations of a long-range wide-body, transonic transport aircraft configuration are performed using the developed techniques. The aerostructural analysis employs Euler CFD with a 2 million cell mesh and a structural finite element model with 300000 DOF. Two design optimization problems are solved: one where takeoff gross weight is minimized, and another where fuel burn is minimized. Each optimization uses a multi-point formulation with 5 cruise conditions and 2 maneuver conditions. The optimization problems have 476 design variables are optimal results are obtained within 36 hours of wall time using 435 processors. The TOGW minimization results in a 4.2% reduction in TOGW with a 6.6% fuel burn reduction, while the fuel burn optimization resulted in a 11.2% fuel burn reduction with no change to the takeoff gross weight.

A parallel, block-based, adaptive mesh refinement, finite-volume scheme is developed and validated for the solution of the Favre-Averaged Navier-Stokes equations governing three-dimensional flow of a polytropic gas. The two-equation k-omega turbulence model is used to model the unresolved turbulent scales and their influence on the mean solution. The finite-volume spatial discretization is accomplished by using a finite-volume procedure on multiblock, body-fitted, hexahedral mesh. The inviscid flux functions make use of Roe's approximate Riemann solver. The viscous flux is evaluated using a diamond path reconstruction procedure on each cell boundary. Verification and validation of the solution method is accomplished through the application of the algorithm to a number of flow problems. The results from the application of the solution method to the flow problems are in good agreement with available experimental data. Therefore, the validity of the solution method for solving three-dimensional, turbulent flows is confirmed.

This thesis presents new tools and techniques developed to address the challenging problem of high-fidelity aerostructural optimization with respect to large numbers of design variables. A new mesh-movement scheme is developed that is both computationally efficient and sufficiently robust to accommodate large geometric design changes and aerostructural deformations. A fully coupled Newton-Krylov method is presented that accelerates the convergence of aerostructural systems and provides a 20% performance improvement over the traditional nonlinear block Gauss-Seidel approach and can handle more flexible structures. A coupled adjoint method is used that efficiently computes derivatives for a gradient-based optimization algorithm. The implementation uses only machine accurate derivative techniques and is verified to yield fully consistent derivatives by comparing against the complex step method. The fully-coupled large-scale coupled adjoint solution method is shown to have 30% better performance than the segregated approach. The parallel scalability of the coupled adjoint technique is demonstrated on an Euler Computational Fluid Dynamics (CFD) model with more than 80 million state variables coupled to a detailed structural finite-element model of the wing with more than 1 million degrees of freedom. Multi-point high-fidelity aerostructural optimizations of a long-range wide-body, transonic transport aircraft configuration are performed using the developed techniques. The aerostructural analysis employs Euler CFD with a 2 million cell mesh and a structural finite element model with 300000 DOF. Two design optimization problems are solved: one where takeoff gross weight is minimized, and another where fuel burn is minimized. Each optimization uses a multi-point formulation with 5 cruise conditions and 2 maneuver conditions. The optimization problems have 476 design variables are optimal results are obtained within 36 hours of wall time using 435 processors. The TOGW minimization results in a 4.2% reduction in TOGW with a 6.6% fuel burn reduction, while the fuel burn optimization resulted in a 11.2% fuel burn reduction with no change to the takeoff gross weight.

A parallel, block-based, adaptive mesh refinement, finite-volume scheme is developed and validated for the solution of the Favre-Averaged Navier-Stokes equations governing three-dimensional flow of a polytropic gas. The two-equation k-omega turbulence model is used to model the unresolved turbulent scales and their influence on the mean solution. The finite-volume spatial discretization is accomplished by using a finite-volume procedure on multiblock, body-fitted, hexahedral mesh. The inviscid flux functions make use of Roe's approximate Riemann solver. The viscous flux is evaluated using a diamond path reconstruction procedure on each cell boundary. Verification and validation of the solution method is accomplished through the application of the algorithm to a number of flow problems. The results from the application of the solution method to the flow problems are in good agreement with available experimental data. Therefore, the validity of the solution method for solving three-dimensional, turbulent flows is confirmed.

The evaluation of the potential hazards related with the operation of an LNG terminal includes possible release scenarios with the consequent flammable vapor dispersion within the facility; therefore, it is important to know the behavior of this phenomenon through the application of advanced simulation tools. Computational Fluid Dynamic (CFD) tools are often used to estimate the exclusion zones in an event of accidental LNG spill. In practice these releases are more likely to occur in the confines of complex geometries with solid obstacles such as LNG terminals, and LNG processing plants. The objective of this research is to study the effects that different obstacles have over the LNG vapor dispersion and the safety distance reduction caused by enhanced mixing. Through parametric analysis it is demonstrated that height, width and shape of the obstacles play an important role in the vapor concentration reduction. The findings of this research may be applied in the design stage of an LNG terminal, to improve the design of passive barriers, and for designing better layout configurations for storage tanks. Simulations results performed with FLACS (Flame Acceleration Simulator), a CFD solver, confirmed that these applications help to reduce safety distances.