Banca de DEFESA: PAULO VITOR DE QUEIROZ FERREIRA

Uma banca de DEFESA de MESTRADO foi cadastrada pelo programa.
STUDENT : PAULO VITOR DE QUEIROZ FERREIRA
DATE: 19/12/2025
TIME: 14:00
LOCAL: remota
TITLE:

A Deep Learning Approach for Time-Lapse Seismic Inversion with Sparse Data

KEY WORDS:

Inverse Problems, 4D Seismic Inversion, Machine Learning, Convolutional Neural Networks (CNNs), Sparse
Data.

PAGES: 81
BIG AREA: Ciências Exatas e da Terra
AREA: Física
SUBÁREA: Física Geral
SPECIALTY: Física Estatística e Termodinâmica
SUMMARY:

Inverse problems are fundamental across a broad scope of physics. Among them, wavefield inversion stands out as a complex,
ill-posed, and often computationally prohibitive inverse problem. The introduction of the temporal dimension (time-lapse) adds
an additional layer of complexity to the inversion, yet it plays an essential role in monitoring the dynamics of regions of
interest, such as oil and gas reservoirs. This work proposes and evaluates the use of convolutional neural networks (CNNs) as a
data-driven and computationally efficient approach to the problem of target-oriented time-lapse wavefield inversion. The
results demonstrate that the proposed methodology is capable of inferring dynamic variations in the region of interest with high
accuracy and at a significantly lower computational cost than conventional methods, presenting itself as a viable alternative for
4D seismic analysis.

COMMITTEE MEMBERS:
Externo à Instituição - BRUNO SOUZA CARMO
Interno - 1379465 - GILBERTO CORSO
Interno - 1328776 - RAFAEL CHAVES SOUTO ARAUJO