| This dissertation presents the theoretical methodology, organizational strategy, conceptual demonstration and validation of a fully automated computer program for the multi-disciplinary analysis, inverse design and optimization of convectively cooled axial gas turbine blades and vanes. Parametric computer models of the three-dimensional cooled turbine blades and vanes were developed, including the automatic generation of discretized computational grids. Several new analysis programs were written and incorporated with existing computational tools to provide computer models of the engine cycle, aero-thermodynamics, heat conduction and thermofluid physics of the internally cooled turbine blades and vanes. A generalized information transfer protocol was developed to provide the automatic mapping of geometric and boundary condition data between the parametric design tool and the numerical analysis programs. A constrained hybrid optimization algorithm controlled the overall operation of the system and guided the multi-disciplinary internal turbine cooling design process towards the objectives and constraints of engine cycle performance, aerodynamic efficiency, cooling effectiveness and turbine blade and vane durability.; Several boundary element computer programs were written to solve the steady-state non-linear heat conduction equation inside the internally cooled and thermal barrier-coated turbine blades and vanes. The boundary element method (BEM) did not require grid generation inside the internally cooled turbine blades and vanes, so the parametric model was very robust. Implicit differentiations of the BEM thermal and thereto-elastic analyses were done to compute design sensitivity derivatives faster and more accurately than via explicit finite differencing. A factor of three savings of computer processing time was realized for two-dimensional thermal optimization problems, and a factor of twenty was obtained for three-dimensional thermal optimization problems, while the program's memory requirements were doubled. This substantial savings was only slightly offset by the fact that a three-dimensional BEM solution was about two to five times slower than a comparable finite element (FEM) analysis.; A compressible thereto-fluid network program was also written to provide a model for an arbitrary network of internal coolant passages for the calculation of the internal heat transfer coefficients, pressure losses, flow rate and coolant fluid heat-up. (Abstract shortened by UMI.)... |
