This week in 4D seismic we have contributions from ASEG extended abstracts and Geophysical prospecting. We also have two special appearances on a statoil publication and bei Marieke van Hout – de Groot of OpenDTect marketing a premium tool on linkedin.
Let’s start with Statoil. They published a multi-scale, multi-phase flow upscaling workflow. I don’t think there is too much news in the publication. However, it gives a great overview of a upscaling workflow in a visually very appealing way. We can also see a busy beautiful representation of the Sleipnir 4D data.
The ASEG abstracts include a couple of personal highlights that I will go into further detail sometime later on this blog (feel free to subsribe!)
Souza, Lumley and Shragge from the Uni of Wester Australia make a case that using 4D data in the stead of inversion products proves to be more reliable to generate water saturation maps. This seems particularly true in the presence of 4D seismic noise. They use a histogram similarity method to quantify their findings on the Namorado Field data benchmark from Brazil.
The paper “Characterizing Heterogeneities in a Clastic Reservoir Using Joint/Simultaneous PP/PS Inversion, 4D Timelapse, Multi Attribute Analysis, and PSDM” by Jason Nycz has a firework of keywoards in its title. Working with two 3D data sets 1 year apart, he tests several inversion projects on the data. The Multi Attribute Analysis (MAT) mentioned in this case is a statistical trace matching algorithm to the borehole logs. This is not exactly my understand of multi-attribute analysis as I figured it would be combining multiple seismic attributes into a common analysis window, but of course it’s quite possible there are several meanings or my understanding is plain wrong. The aforementioned heterogeneities are the main reason, borehole derived velocities, PSTM, MAT fail to deliver reliable results. The 4D difference however validates missing communication of some injectors and producers due to siderite layers.
Lahra Lanigan and Jarrod Dunne of Karoon Gas Australia Ltd have ventured into the realms of bridging the scale between core scale, wireline logs and possibly seismic data. However, their AI-Vp/Vs AVO response of derived core velocities showed very small differences that were easily overridden by noise or variation of volume in the carbonates. The main reason they suggest is a very close similarity of the sands that were included in this research. Overall it seems there may be more fuel behind this study but since the differences are usually very small it takes good circumstances and some fine-tuning to get reproducible results.
In the article “Sub-sample time shift and horizontal displacement measurements using phase-correlation method in time-lapse seismic”, published in Geophysical prospecting ,Tomar et al. propose an application of a phase correlation method that was developed for satellite imaging in subpixel resolution. It is tested on Norne field data. I have yet to read the entire article as I don’t have remote access to it. However, I find the idea compelling.
Marieke van Hout – de Groot is the international sales manager at dGB Earth Sciences. It’s quite interesting to follow her on linkedin, as she often posts insights into Open dTect. This time the article is about “Enhancing 4D Anomalies with OpendTect Pro’s Local Fluid Contact Finder“. The Fluid Contact Finder is a premium plugin to locate structural traps in 3D seismic data sets. The usual approach of the FCF module would eliminate any 4D differences due to stacking. However, they introduced a local variety that let’s the user provide a enhancement radius. Honestly, I would be extremely careful, applying this technique to 4D differences. It may be valid in some cases but I can definitely see how it might do more harm than good if trusted blindly.
She links to this presentation, which is nice, but I do not see the value for 4D. It seems like an attempt to gain a wider userbase without supplying the appropriate tools, hacking existing tools to losely make a usecase. I would love to see some genuinely 4D implementations in Open dTect, as I like the open source aspect of it.
Latest posts by Jesper Dramsch (see all)
- Research Talk — Deep Learning for 4D Pressure Saturation Inversion [Youtube] - 2019-04-18
- Geysers in Slow Motion - 2019-02-04
- Keynote Bonanza and No Coffee – The EAGE / PESGB ML Workshop - 2018-12-17