2020/11/12 by Chaitanya Talnikar, Talnikar, Chaitanya, Qiqi Wang +1
Computer Science · Engineering · #Advanced Multi-Objective Optimization Algorithms #FOS: Physical sciences #Fluid Dynamics (physics.flu-dyn) #Heat Transfer Mechanisms #Turbomachinery Performance and Optimization
paper · pdf · doi:10.48550/arxiv.2011.06744
openalex publication_date 2020/11/12 · openalex created_date 2022/07/25 · openalex updated_date 2026/07/28
The shape of the trailing edge of a gas turbine nozzle guide vane has a\nsignificant effect on the downstream stagnation pressure loss and heat transfer\nover the surface of the vane. Traditionally, adjoint-based design optimization\nmethods for turbomachinery components have used low-fidelity simulations like\nReynolds averaged Navier-Stokes. To reliably capture the complex flow phenomena\ninvolved in turbulent flow over a turbine vane, high-fidelity simulations like\nlarge eddy simulation (LES) are required. In this paper, an adjoint-based\ntrailing edge shape optimization using LES is performed to reduce pressure loss\nand heat transfer over the surface of the vane. The chaotic dynamics of\nturbulence limits the effectiveness of the adjoint method for long-time\naveraged objective functions computed from LES. A viscosity stabilized unsteady\nadjoint method is used to obtain gradients of the design objective function\nwith reasonable accuracy. A gradient utilizing Bayesian optimization is used to\nrobustly handle noise in the objective function and gradient evaluations. The\ntrailing edge shape is parameterized using a linear combination of 5 convex\ndesigns. Results from the optimization, performed on the supercomputer Mira,\nare compared with optimal designs generated using derivative-free design\noptimization of the same problem.\n