Banca de DEFESA: LEANDRO ROBERTO SILVA FARIAS

Uma banca de DEFESA de MESTRADO foi cadastrada pelo programa.
STUDENT : LEANDRO ROBERTO SILVA FARIAS
DATE: 03/07/2026
TIME: 10:00
LOCAL: Sala virtual do Google Meet
TITLE:

Seismic Q-Factor Inversion Strategy via Non-Stationary Convolutional Modeling

KEY WORDS:

quality factor Q; seismic attenuation; non-stationary convolutional model; deconvolution; gradient descent.

PAGES: 69
BIG AREA: Engenharias
AREA: Engenharia Elétrica
SUMMARY:

Estimating the seismic quality factor Q is an important problem in seismic processing, since attenuation directly affects the amplitude, phase, and resolution of the propagated signal. Classical Q-estimation methods are generally better suited for VSP data and usually treat attenuation separately from the deconvolution stage. This work proposes a Q-factor estimation strategy based on a non-stationary convolutional model, in which dissipation and dispersion effects are explicitly parameterized by the quality factor Q. The methodology is formulated as a minimization problem between observed and modeled seismic data and solved using gradient-based optimization methods. Supervised and blind scenarios are considered, together with analytical derivations of the cost-function gradient. Preliminary results, obtained from a synthetic seismic trace generated with the wave equation, show that the proposed approach was able to recover the main structure of the Q profile in both the supervised and blind cases, indicating its potential for more realistic applications.

COMMITTEE MEMBERS:
Presidente - 3216921 - TIAGO TAVARES LEITE BARROS
Interno - 1543191 - LUIZ FELIPE DE QUEIROZ SILVEIRA
Externo à Instituição - KENJI NOSE FILHO - UFABC