New Modified Band Limited Impedance (BLIMP) Inversion Method Using Envelope Attribute

Padjadjaran Earth Dialogues: International Symposium on Geophysical Issues

Maulana, Z.L.^1 ; Saputro, O.D.^1 ; Latief, F.D.E.^1

Department of Physics, Faculty of Mathematics and Natural Sciences, Institut Teknologi Bandung, Jalan Ganesa 10, Bandung

40132, Indonesia^1

关键词: Conventional methods;    Low frequency domain;    Mean squared error;    Normalization constants;    Reservoir property;    Seismic attributes;    Synthetic seismogram;    Velocity analysis;   

Others  :  https://iopscience.iop.org/article/10.1088/1755-1315/29/1/012009/pdf
DOI  :  10.1088/1755-1315/29/1/012009

来源: IOP

【 摘 要 】

Earth attenuates high frequencies from seismic wavelet. Low frequency seismics cannot be obtained by low quality geophone. The low frequencies (0-10 Hz) that are not present in seismic data are important to obtain a good result in acoustic impedance (AI) inversion. AI is important to determine reservoir quality by converting AI to reservoir properties like porosity, permeability and water saturation. The low frequencies can be supplied from impedance log (AI logs), velocity analysis, and from the combination of both data. In this study, we propose that the low frequencies could be obtained from the envelope seismic attribute. This new proposed method is essentially a modified BLIMP (Band Limited Impedance) inversion method, in which the AI logs for BLIMP substituted with the envelope attribute. In low frequency domain (0-10 Hz), the envelope attribute produces high amplitude. This low frequency from the envelope attribute is utilized to replace low frequency from AI logs in BLIMP. Linear trend in this method is acquired from the AI logs. In this study, the method is applied on synthetic seismograms created from impedance log from well 'X'. The mean squared error from the modified BLIMP inversion is 2-4% for each trace (variation in error is caused by different normalization constant), lower than the conventional BLIMP inversion which produces error of 8%. The new method is also applied on Marmousi2 dataset and show promising result. The modified BLIMP inversion result from Marmousi2 by using one log AI is better than the one produced from the conventional method.