Shape optimization of inclined hole for enhanced film-cooling performance using discrete adjoint method
| Autor: | Jiazhen Leng, Aziz Madrane, Damiano Pasini, Minh Quan Pham, Ali Shanian, Megan Schaenzer, Genevieve Bourgeois, Haichao An |
|---|---|
| Rok vydání: | 2020 |
| Předmět: |
Physics
Bearing (mechanical) business.industry 020209 energy General Engineering Boundary (topology) 02 engineering and technology Mechanics Computational fluid dynamics Condensed Matter Physics 01 natural sciences Symmetry (physics) 010305 fluids & plasmas law.invention law 0103 physical sciences Line (geometry) 0202 electrical engineering electronic engineering information engineering Shape optimization Asymptote Reynolds-averaged Navier–Stokes equations business |
| Zdroj: | International Journal of Thermal Sciences. 158:106542 |
| ISSN: | 1290-0729 |
| Popis: | This paper presents a gradient-based framework to optimize the film-cooling performance of heat bearing components in gas turbines. We examine a representative periodic component with very low porosity (below 1 % ) and optimize the shape of an inclined hole on the wall of a combustor for maximum cooling effectiveness, here defined as the normalized average decrease of downstream temperature. The shape boundary of the hole is parameterized using the coordinates of its control points here describing the design variables. Geometric constraints are introduced on the void area, symmetry relations and relative motion between planes along the line of extrusion. The cooling effectiveness is numerically evaluated through a RANS (Reynolds-Averaged Navier-Stokes)-based computational fluid dynamics (CFD) analysis. The Method of Moving Asymptotes (MMA) is adopted as the optimizer, with gradients computed using the adjoint method. Nine case studies with initial hole geometry of circular and elliptical shape as well as various inclination angles converge to a family of V-shapes that achieve more than 120 % increase in cooling effectiveness. The automation of the workflow presented in this paper can be implemented in CFD-based optimization of other single and multi-objectives problems. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|
Full Text from ScienceDirect