Optimization based on the Adjoint method for adsorbed natural gas storage systems.

• Autor(a) principal: Chieregatti, Bruno Galelli
• Data de Publicação: 2020
• Tipo de documento: Tese
• Idioma: eng
• Título da fonte: Biblioteca Digital de Teses e Dissertações da USP
• Texto Completo: https://www.teses.usp.br/teses/disponiveis/3/3150/tde-15122020-100537/
• topic: Adjoint method; Adsorbed natural gas; CFD; Dinâmica dos fluídos computacional; Gás natural (Armazenagem; Sistemas); Gestão por processos (Otimização); Método adjunto; Optimization

This PhD thesis aims to develop and investigate the application of the socalled Adjoint Method in flows through porous media. Its main focus is on the Adsorbed Natural Gas (ANG) Storage Systems, a growing concept in Natural Gas equipments, but the formulation developed is applied in all porous media flows, considering (or not) the Adsorption Phenomena. The primary objective is to optimize their filling process performance which consists an adsorption process. To that end, it shall consider controlling not only geometry parameters, as the tank dimensions, but also the non geometric parameters, such as filling flow curves, temperature fields and heat transfer coefficients. These kinds of control devices and strategies have their niche in small and specially in large scale systems that can be found in power and industrial plants. Owing to their strong dependence on both the system geometry and on the thermodynamics of the adsorption processes, this class of applications could greatly benefit from parametric and form optimization techniques. That is precisely the rationale behind the choice of the Adjoint Method, which can in principle serve both purposes. In that regard, it should be added that, although the physics of the adsorption processes is well documented in the literature, there seems to be very few references that consider their optimization, and none that make use of the AdjointMethod. Under such circumstances, t his thesis developed a strong mathematical formulation, starting from the basic equations of fluid mechanics, where applying the suitable hypothesis, the physics flow were been modeled and validated. The Adjoint Equations received the same treatment, starting from the Lagrange Multipliers until the study of the Adjoint Contour Problem. The results, not only produces values of the sensitivity gradients of some objective functions but also present a dramatically reduction of computational cost, in compassion between a classic method, called Central Finite Difference. A study of an optimization of a filling flow curve is done in the end of the work, showing the possibity of the use this tool in engineering problems.