Multi-objective design optimization of turbine blade leading edge for enhanced aerothermal performance

A multi-objective design optimization is carried out to minimize the heat transfer in the leading edge region of an uncooled turbine blade and the blade profile loss. These objectives pose conflicting requirements. The leading edge of Pratt and Whitney JT9D turbine blade is parameterized by Beizer curves. The Latin hypercube sampling plan is used to sample the design space. Several turbine blade geometries are created and their heat transfer and aerodynamic characteristics are evaluated using high-fidelity Reynolds-averaged Navier-Stokes (RANS) simulations. Kriging surrogate models are constructed using these datasets. The surrogate models are used in a genetic algorithm optimization framework to obtain optimal designs. The maximum heat transfer in the leading edge region is reduced by 5.9%. The blade profile loss is reduced by 18.1%. The surrogate models are then subject to multi-objective genetic algorithm optimization to reduce both the maximum heat transfer and the blade profile loss. A Pareto-optimal front is obtained which contains the optimal solutions. Some optimal solutions in the Pareto front are chosen and the trade-off between the competing objective functions is presented.