A General Multidisciplinary Turbomachinery Design Optimization system Applied to a Transonic Fan

The blade geometry design process is integral to the development and advancement of compressors and turbines in gas generators or aeroengines. A new airfoil section design capability has been added to an open source parametric 3D blade design tool. Curvature of the meanline is controlled using B-splines to create the airfoils. The curvature is analytically integrated to derive the angles and the meanline is obtained by integrating the angles. A smooth thickness distribution is then added to the airfoil to guarantee a smooth shape while maintaining a prescribed thickness distribution. A leading edge B-spline definition has also been implemented to achieve customized airfoil leading edges which guarantees smoothness with parametric eccentricity and droop.;An automated turbomachinery design and optimization system has been created. An existing splittered transonic fan is used as a test and reference case. This design was more general than a conventional design to have access to the other design methodology. The whole mechanical and aerodynamic design loops are automated for the optimization process. The flow path and the geometrical properties of the rotor are initially created using the axi-symmetric design and analysis code (T-AXI). The main and splitter blades are parametrically designed with the created geometry builder (3DBGB) using the new added features (curvature technique). The solid model creation of the rotor sector with a periodic boundaries combining the main blade and splitter is done using MATLAB code directly connected to SolidWorks including the hub, fillets and tip clearance. A mechanical optimization is performed with DAKOTA (developed by DOE) to reduce the mass of the blades while keeping maximum stress as a constraint with a safety factor. A Genetic algorithm followed by Numerical Gradient optimization strategies are used in the mechanical optimization. The splittered transonic fan blades mass is reduced by 2.6% while constraining the maximum stress below 50% material yield strength using 2D sections thickness and chord multipliers.;Once the initial design was mechanically optimized, a CFD optimization was performed to maximize efficiency and/or stall margin. The CFD grid generator (AUTOGRID) reads 3DBGB output and accounts for hub fillets and tip gaps. Single and Multi-objective Genetic Algorithm (SOGA, MOGA) optimization have been used with the CFD analysis system. In SOGA optimization, efficiency was increased by 3.525% from 78.364% to 81.889% while only changing 4 design parameters. For MOGA optimization with higher weighting efficiency than stall margin, the efficiency was increased by 2.651% from 78.364% to 81.015% while the static pressure recovery factor was increased from 0.37407 to 0.4812286 that consequently increases the stall margin.;The design process starts with a hot shape design, and then a hot to cold transformation process is explained once the optimization process ends which smoothly subtracts the mechanical deflections from the hot shape. This transformation ensures an accurate tip clearance.;The optimization modules can be customized by the user as one full optimization or multiple small ones. This allows the designer not to be eliminated from the design loop which helps in taking the right choice of parameters for the optimization and the final feasible design.