|
" Estimation of Reservoir Properties from Seismic Attributes and Well Log Data using Artificial Intelligence "
Mohamed Sitouah
G. A. O. Korvin, Abdul Latif
Document Type
|
:
|
Latin Dissertation
|
Language of Document
|
:
|
English
|
Record Number
|
:
|
54769
|
Doc. No
|
:
|
TL24723
|
Call number
|
:
|
1533986
|
Main Entry
|
:
|
Mohamed Sitouah
|
Title & Author
|
:
|
Estimation of Reservoir Properties from Seismic Attributes and Well Log Data using Artificial Intelligence\ Mohamed Sitouah
|
College
|
:
|
King Fahd University of Petroleum and Minerals (Saudi Arabia)
|
Date
|
:
|
2009
|
Degree
|
:
|
M.S.
|
student score
|
:
|
2009
|
Page No
|
:
|
156
|
Abstract
|
:
|
Permeability, Porosity and Lithofacies are key factors in reservoir characterizations. Permeability, or flow capacity, is the ability of porous rocks to transmit fluids, porosity, represent the capacity of the rock to store the fluids, while lithofacies, describe the physical properties of rocks including texture, mineralogy and grain size. Many empirical approaches, such as linear/non-linear regression or graphical techniques. Were developed for predicting porosity, permeability and lithofacies. Recently, researches used another tool named Artificial Neural Networks (ANNs) to achieve better predictions. To demonstrate the usefulness of Artificial Intelligence technique in geoscience area, we describe and compare two types of Neural Networks named Multilayer Perception Neural Network (MLP) with back propagation algorithm and General Regression Neural Network (GRNN), in prediction reservoir properties from seismic attributes and well log data. This study explores the capability of both paradigms, as automatique systems for predicting sandstone reservoir properties, in vertical and spatial directions. As it was expected, these computational intelligence approaches overcome the weakness of the standard regression techniques. Generally, the results show that the performances of General Regression neural networks outperform that of Multilayer Perceptron neural networks. In addition, General Regression Neural networks are more robust, easier and quicker to train. Therefore, we believe that the use of these better techniques will be valuable for Geoscientists.
|
Subject
|
:
|
Earth sciences; Geophysics; 0373:Geophysics
|
Added Entry
|
:
|
G. A. O. Korvin, Abdul Latif
|
Added Entry
|
:
|
King Fahd University of Petroleum and Minerals (Saudi Arabia)
|
| |