Navegando por Assunto "Redes neurais - Computação"

Agora exibindo 1 - 4 de 4

Acesso aberto (Open Access)
Aplicação de redes neuronais artificiais ao tratamento e interpretação de perfis geofísicos de poço aberto
(Universidade Federal do Pará, 1997-02-13) ANDRADE, André José Neves; LUTHI, Stefan Moritz
The analysis of openhole wireline logs is of great importance for the subsurface mapping of geological layers and the identification and quantification of hydrocarbon and mineral deposits. An importants aspects are the determination of geological boundaries, the mineralogical identification and the well-to-well correlation, which can be a tedious and time-consuming task for the geologist. Automating this procedure is complicated but potentially rewarding because it may save the production geologist and log analyst substantial amounts of time. Artificial neural networks have been shown to handle this task efficiently including in cases where sequential algorithms have problems. We show in this thesis that neural networks can be used to determine layer boundaries, the mineralogical identification and the well-to-well correlation, on well logs, and we present the new networks archtectures. These procedures are tested on synthetic as well as actual field data.
Acesso aberto (Open Access)
Cálculo da porosidade: identificação do argilomineral
(Universidade Federal do Pará, 2017-04-20) ALMEIDA, Thales Luiz Pinheiro de; ANDRADE, André José Neves; http://lattes.cnpq.br/8388930487104926
In the daily practice of Formation Evaluation it is common the adoption of simplifying hypotheses or premises about the physical properties of the reservoir rock constituent materials to allow the porosity calculation. The knowledge of the physical properties of the clay in the reservoir rock is fundamental for porosity calculation. In this work it is argued that the physical properties of the clay mineral in the reservoir rock are different from the mean physical properties of the nearby shale layers. Geologically, the clay is one of the constituent materials of shale rock and to admit that the physical properties of the shale are equal to the physical properties of the clay in the reservoir rock means to disregard all the other constituents and to assume a sedimentary continuity that due to numerous postpositional processes may not occur. In this work, we apply the angular competitive neural network to the Density- Neutron Plot to show that if a reservoir rock and a shale present in the basin have the same clay, they have the same angular pattern. This methodology is presented with synthetic data and evaluated with actual well logs and core analysis from borehole drilled in the Namorado’s field, in the Campos Basin, Brazil.
Acesso aberto (Open Access)
Correlação de poços com múltiplos perfis através da rede neural multicamadas
(Universidade Federal do Pará, 2001-11-23) AMARAL, Mádio da Silva; ANDRADE, André José Neves; http://lattes.cnpq.br/8388930487104926
Stratigraphic correlation using well logs is a non-trivial geological activity and subject to endless possibilities of misunderstanding about the geometry or continuity of rock layers, for many reasons, like the geological variability and the ambiguous answers of the log tools. Thus, it is common to utilize a great log suite from the same well, for better comprehension. The stratigraphic correlation is a fundamental tool for a geologist or petroleum geophysist, because from its knowledge it is possible to interpret the hydraulic continuities of the reservoirs and to reconstruct the geological setting environment, which may corroborate for the construction of the reservoir geological model. This work produces an automation of manual activities involved in the stratigraphic correlation, with the use of the various well logs, and a convenient architecture of artificial neural network, trained with the backpropagation algorithm. The stratigraphic correlation, obtained from this method, makes the transport of the geological information possible along the basin and gives the interpreter, a general view of the structural behavior of the oil reservoir. With This methodology was possible the automatic construction of a geological block diagram showing the spatial disposition of a particular shale layer, from the well logs: Gamma Ray (GR), Clay Volume (Vsh), Density (ρb) and the Neutron Porosity (φn), selected in the five wells on the Maracaibo Lake basin, in Venezuela.
Acesso aberto (Open Access)
Identificação de fácies em perfis de poço com algoritmo inteligente
(Universidade Federal do Pará, 2014-02-17) SANTOS, Renata de Sena; ANDRADE, André José Neves; http://lattes.cnpq.br/8388930487104926
Facies identification in an uncored borehole is a classic problem in formation evaluation. In this study, this problem is treated as the extraction of geological information or facies descriptions from a cored borehole in terms of their physical properties registered in well logs and perform their encoding through the parameters L and K calculated from porosity logs, and shaliness calculated using the natural gamma ray log to construct the Vsh-L-K plot. For interpretation is presented an intelligent algorithm based on the competitive generalized angular neural network, built for angular pattern classification or data clustering in ndimensional space that have an approximately ellipsoidal envelope, which are the characteristics of clusters in the Vsh-L-K plot and make your visual interpretation extremely complex. The application of intelligent algorithm is able to identify and classify the layers present in uncored boreholes, in terms of the facies identified in the cored borehole or in terms of its main mineral, when it is absent in the cored borehole. This methodology is presented with synthetic data and well logs from cored boreholes in Namorado oil field, in the Campos Basin, located on the continental shelf of Rio de Janeiro, Brazil.