Banca de DEFESA: ADRIANO DAYVSON MARQUES FERREIRA

Uma banca de DEFESA de DOUTORADO foi cadastrada pelo programa.
STUDENT : ADRIANO DAYVSON MARQUES FERREIRA
DATE: 30/08/2022
TIME: 09:00
LOCAL: Programa de Pós-Graduação em Engenharia Civil
TITLE:

Multiresolution Analysis and Deep Learning for Corroded Pipeline Failure Assessment

KEY WORDS:

Corroded Pipelines, Multiresolution analysis, Discrete Wavelet Transform, Deep Neural Network, Finite Element Method

PAGES: 154
BIG AREA: Engenharias
AREA: Engenharia Civil
SUMMARY:

The structural integrity of corroded pipelines is considered a vital task in the oil and gas industry. This research aims to develop an efficient system to accurately predict the burst pressure of corroded pipelines with complex corrosion profiles through hybrid models using multiresolution analysis, numerical analysis, and metamodels. These work addresses the parametrization of real corrosion shapes and its use as input to a neural network systemthat can accurately predict the burst pressure quickly. The corrosion map is obtained from ultrasonic inspections and the data is used both in the form of a river bottom profile and in the form of a three-dimensional mapping. The finite element method (FEM) is used to evaluate the burst pressure. Scripts to automatically generate axisymmetric and three-dimensional finite element (FE) models are used and failure pressures are obtained with non-linear analysis. The FE models and analysis procedure are validated against experimental tests and are compared with semi-empirical assessment methods. A discrete wavelet transform is performed for the parametrization of the remaining thicknesses and as a filter bank to reduce the amount of data do describes the defect. The coefficients obtained from the wavelet transform and the material properties of the pipelines are used as inputs to feed a deep neural network. Axisymetric and three-dimensional synthetic models that have similar statistics to real corrosion profiles are created and submitted to non-linear FEM analysis. The respective failure pressures obtained from the synthetic defects are used to train a neural network to predict the burst pressure of the pipelines with river bottom profile and to train a neural network to predict the burst pressure of pipelines with real corrosion defects. The results obtained with the deep neural networks are very precise for all the cases presented in this work, both in the use of axisymmetric models and in the use of three-dimensional models.

BANKING MEMBERS:
Externo à Instituição - BRENO PINHEIRO JACOB - UFRJ
Interno - 1130915 - BERNARDO HOROWITZ
Externo à Instituição - JOSE ANGELO PEIXOTO DA COSTA - IFPE
Externo ao Programa - 1329483 - PAULO FERNANDO SILVA SOUSA
Presidente - 2250444 - RENATO DE SIQUEIRA MOTTA