LaMEP qualifies doctoral student for thesis in deep learning applied to oil industry

/ December 19, 2018/ Thesis defenses

On last December 7th, Rafael Magalhães, researcher at LaMEP and current doctoral student at Postgraduate Program in Mechanical Engineering/UFPB, was qualified for his doctoral thesis. With a background in Computer Science, Rafael began to developing codes for deep learning training in the past, but now he conciliates his knowledge to reservoir modelling.

Rafael’s work plan, entitled Deep Learning-Based Algorithms as Proxy Models for Application in Hydrocarbon Reservoirs, stands on the main rationale of using surrogate models to solve a large amount of problems in reservoir engineering. According to the author, the high complexity of deep learning architectures associated to its myriad of free parameters are able to mimic the real behaviour of several phenomena occurring in an oilfield. Moreover, he says that from a suitable training process, similar results to those achieved through conventional and proprietary simulators regarding management and production can be obtained from artificial neural networks.

The Examining Board for Rafael’s thesis was formed by experts in Petroleum Engineering, Mechanical Engineering and Computing, namely Profs. Moisés Dantas and Gustavo Oliveira (LaMEP), Abel Lima and Marcelo Cavalcanti (PPGEM/UFPB), and Gilberto Farias (PPGI/UFPB), whereby he was unanismouly approved. The Board has recalled that Rafael is already well prepared in intelligent systems and teaching activities. His proposal was praised due to its relevance for oil industry in the ambit of digital and intelligent oilfields (i-fields). In 2019, Rafael will go abroad for a 1-year internship in Grenoble, France.

See below some photographies of the event.

Rafael talks about challenges for the Industry 4.0 wave related to digital oilfields.
Examining board: Profs. Abel Lima, Gustavo Oliveira, Gilberto Farias, Moisés Dantas and Marcelo Cavalcanti (from left to right).
Share this Post