Topology optimization of compressible subsonic flow considering turbulence and real gases.

• Ano de defesa: 2022
• Autor(a) principal: Garcia Rodriguez, Luis Fernando
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: eng
• Instituição de defesa: Biblioteca Digitais de Teses e Dissertações da USP
• Palavras-chave em Português: Compressibilidade do fluxo dos fluidos; Compressibilidade dos gases; Compressible flow; FEM; FEniCS; FVM; Métodos topológicos; OpenFoam; Optimization; Otimização matemática; perfect gas; Real gas; Topology; Turbulence; Turbulência
• Link de acesso: https://www.teses.usp.br/teses/disponiveis/3/3152/tde-23052023-080153/

The optimization of compressible turbulent flow is fundamental to improve the performance of turbo-machinery and fluid devices used at cutting-edge energy transition technologies such as the carbon, capture and storage industry. From the different fluid flow optimization methods, the topology optimization offers a flexibility and robustness in the design as handles different objective functions and independence on initial domains that helps to create innovative and non-intuitive domains with splitters and fluid directors. The topology optimization method is performed by changing the permeability of the domain cells through the characterization of the fluid flow through CFD modelling, which is coupled to optimization methods based on the calculation of derivatives acquisition of a Lagrange system. To guarantee a successful topology optimization stable CFD models to be employed, which for compressible turbulent flow are available at subsonic regime according to the literature review. Therefore, the current research develops the topology optimization formulation for compressible turbulent subsonic flow, which demands the use of compressibility effects, rotational influence, and turbulence phenomenon besides the acquisition of the compressible adjoint system. The methodology considers the finite volume method, and the proposed formulation establishes the minimization of the energy dissipation subjected to the compressible Favre Averaged Navier-Stokes equations, the permeability of the volume cells and a volume constraint. The acquisition of the compressible adjoint system is studied simultaneously at two different approaches: via the continuous forward system and the discrete adjointautomatic differentiator, the last adapted and coupled to finite volume method applications. Also, the influence of the integer design variables-based optimizer is included, which presents a considerable advantage over the common continuous design variables optimizers used. Finally, the optimization considering real gas modelling for compressible turbulent subsonic flow is presented, where the compressibility and temperature effects are embraced by the Peng-Robinson and the Sutherlands law. Different tests are performed in channels and rotors that validate the mentioned hypothesis in 2D and 3D domains.