Отправлено: 30.08.17 06:41. Заголовок: Kpr-NMR_NN (Google отыскал)

Inversion of the permeability of a tight gas reservoir with
the combination of a deep Boltzmann kernel extreme
learning machine and nuclear magnetic resonance logging
transverse relaxation time spectrum data

Zhu, Linqi
Zhang, Chong
Wei, Yang
Zhou, Xueqing
Huang, Yuyang
Zhang, Chaomo

Interpretation-2016

Abstract
In view of the low accuracy of the existing NMR logging permeability model in tig
ht sandstone reservoirs, we derive a relationship between the nuclear magnetic resonance
T2 spectrum and permeability based on the transverse relaxation theory of nuclear magnet
ic resonance and the Kozeny-Carman equation. We determined the reasons for the low a
ccuracy of the model through the theoretical analysis. We propose the deep Boltzmann k
ernel extreme learning machine to improve the deep learning algorithm and to predict the
reservoir permeability based on nuclear magnetic resonance logging with a deep Boltzm
ann machine. We use the permeability data of 200 rock specimens in a tight gas reservo
ir in a certain area and the corresponding T2 spectra from NMR logging for modeling.
We apply the model to the evaluation of permeability in this area. The results show that
the accuracy of the deep learning algorithm is higher than that of the existing NMR lo
gging permeability model and the shallow layer machine learning model. Further, the acc
uracy of the deep Boltzmann kernel extreme learning machine proposed by this paper is
higher than that of the deep Boltzmann machine, which indicates that deep Boltzmann ke
rnel extreme learning machine is more suitable for the prediction of reservoir permeabilit
y. Therefore, deep learning theory can be effectively used in oil exploration and develop
ment, it can improve the interpretation accuracy of reservoir parameters. These findings c
ontribute to the interpretation of reservoir parameters.