Desenvolvimento e validação de simulador de tomógrafo óptico para escoamentos bifásicos

• Ano de defesa: 2021
• Autor(a) principal: Bernardelli, Rafael Sturaro
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Tecnológica Federal do Paraná; Curitiba; Brasil; Programa de Pós-Graduação em Engenharia Elétrica e Informática Industrial; UTFPR
• Palavras-chave em Português: Escoamento bifásico; Tomografia óptica; Métodos de simulação; Redes neurais (Computação); Algorítmos genéticos; Two-phase flow; Optical tomography; Simulation methods; Neural networks (Computer science); Genetic algorithms; CNPQ::ENGENHARIAS::ENGENHARIA ELETRICA::ELETRONICA INDUSTRIAL, SISTEMAS E CONTROLES ELETRONICOS::AUTOMACAO ELETRONICA DE PROCESSOS ELETRICOS E INDUSTRIAIS; Engenharia Elétrica
• Link de acesso: http://repositorio.utfpr.edu.br/jspui/handle/1/28828

Two-phase flows are found in several industrial processes, such as oil and gas exploration and production, in thermoelectric plants and in chemical reactors. This motivates the development of technologies for sensing, control and characterization of those phenomena. Tomography is a class of sensing technologies that aims to image cross sections of two-phase flows. Among them, optical tomography stands out because of its low cost, high temporal resolution and mainly because it is an non-intrusive technique. However, this technique calls for a reconstruction method, in order to turn the sensor readouts into cross-sectional images of the flow. The techniques traditionally used for optical tomography exhibit limitations, and this is the motivation for the development of new reconstruction techniques. The employment of simulated optical tomography sensing as the basis for a new method is on the horizon of this research. In this sense, this work aims at the development and validation of the accurate simulation of the light transport phenomenon that takes place inside the optical tomography system coupled with the air–water flow, the object of sensing. The synthetic data generated by the simulation can be used to train new reconstruction models. Alternatively, the simulator itself can be embedded in a reconstruction algorithm. In detail, this work makes progress on the numeric validation for the simulator using real experiments as reference, in addition to advancing on the development of the simulation software. The simulator is GPU-accelerated, and it uses Path Tracing, a Monte Carlo method for light transport simulation. Using an infrared optical tomography system, measurements of phantoms as well as measurements of air–water flows in a bubble column were performed. The numerical validation for the simulator was carried out by comparing the measurements with their respective simulations. As reference for the simulator, three-dimensional models were built based on the geometry of the phantoms. Subsequently, the three-dimensional geometry of the air–water flow was generated using measurements from a Wire-Mesh sensor placed downstream the optical tomography system. The results based on the phantoms proved to be satisfactory, with PSNR values above 20 dB, whereas the tests made with measured flow data were inconclusive. The results of the flow experiments could not be evaluated. This is due to the fact that the deformation of the air–water interfacial surfaces on the way from the tomography sensor to the Wire-Mesh sensor rendered the numeric comparisons not viable. Despite that, the overall results encourage the adoption of the presented simulator to support the development of advanced reconstruction algorithms in an effort to overcome the limitations presented by the usage of traditional tomography reconstruction methods when applied to the optical tomography of two-phase flows.