[ytlee]

This discussion arised from my other thread http://forum.petro-c...0&#entry102011.

In the late 70s, I worked as offshore manager for a company where Ken Lay was the Big Chess . I faced a major problem because I knew nothing about seismic. I couldn't communicate with my GG staff. To over come my problem, I got a seismic guru from Exxon helping me out.

One day while I was watching my favor PBS show, a show of people digging up dinosaur bones. A guy was using a brush removing the dirt. As the dirt was peeling off layer by layer, the outline of the bone was slowly showing up. A crazy idea popped up my head, why can't I use the same technique to peel away the layer of unwanted seismic attribute. I talked to my guru about my idea next day. He, being a guru agreed that was a crazy idea and it wouldn't give any meaningful result. However, he was willing to give it a try. We created an artificial dataset. To our surprise, the crazy idea worked. Unfortunately, he died of a heart attrack playing basket ball at a young age of 40. and I did not get another "guru" to try my idea until I learned how to visualize seismic data myself. I was a slow learner and took 10 years before I got a movie showing the layer of seismic dataset peeling off layer by layer. It took another 10 years before this technique used by the industry, now every commercial package has this feature.

I dug up the movie from my storage shack but I can't upload it because of the file size.

I will upload the movie when the forum admin show me how to upload a 18M movie to the forum

[wzgme]

Oh, this is seismic volume rendering, or no? Transparency sculpting is a popular technology. Kidd introduced this technology in 1999. I am looking forward to your video which will give us more detail.

Now, our team try to integration of simulation and seismic attribute. We design this two-way workflow:

History matching->Gird model->Synthetic seismic attributes

Seismic Data->Attributes->Gird model->History matching

[ytlee]

引用框（wzgme ＠ 12-22-2010 - 06:21）

Oh, this is seismic volume rendering, or no? Transparency sculpting is a popular technology. Kidd introduced this technology in 1999. I am looking forward to your video which will give us more detail.

Now, our team try to integration of simulation and seismic attribute. We design this two-way workflow:

History matching->Gird model->Synthetic seismic attributes

Seismic Data->Attributes->Gird model->History matching

Hello wzgme:

I called it "CLOUD CHAMBER" when I showed it to a GeoQuest colleague.

If you allowed me, I liked to tell a story about my GeoQuest colleague. I met him when I was working on a project for their consulting firm JR Butler. The firm shared an office building with GeoQuest in Post Oak Place. My colleage was a high school student back then. He was working in programming. We met in the computer room and started with idle small talk. He gave me an impression that he is a very smart person and loves what he was doing. I met him again 20 years later at San Felipe campus in late 90s. He was still with GeoQuest but GeoQuest is part of big blue. He was THE EXPERT of Geoframe. Unfortunately, because he never get around to get a college sheepskin, he was struck in a lowly position. He and I co-operated on porting the Geoframe from mainframe to PC. We proved to the big Chess that Geoframe can run on lowly PC. However, we won the battle but lose the war. Big blue didn't want PC to destroy their mainframe business.I left big blue and lose the chance for Petrel. Petrel came out two years later on a PC platform.

Any as for your two-way workflow. If it is not against your company NDA, I would love to see what your team doing before I make meaningful comments. In general, I believe AHM is dangereous if human is not in the loop. I'm not a fan of mouse clicking nor black box/feed the duck club. I'm not for AHM not because I'm old school. I'm against it because it gave management/user/client/etc a false sense of integrity of the study. Users tend to abuse or mis-use the AHM technology. Anyway, I feel your two-way workflow may be a good fit for monitoring reservoir fluid using 4D seismic assuming all reservoir properties changes are result of fluid withdrawal. I would love to see your technology to determine if it is a fit for an OBS project we are planning for a deepwater field.

YT

[ytlee]

part 1 of 3 files

THe seismic dataset is a 3D set of a field in New Zealand. We were trying to visualize the field fault.

This movie shows the outline of the fault as moving down by depth.

附加文件

•  peeling.part1.rar (7.15MB)
  下载次数: 44

[ytlee]

part 2 & 3 of 3 files

附加文件

•  peeling.part2.rar (7.15MB)
  下载次数: 37
•  peeling.part3.rar (711.02K)
  下载次数: 36

[wzgme]

Hello ytlee:

As a new guy in Geophysics, I am sorry that I never use GeoQuest. I have watched your movie many times, and I already have a intuitive impression. If I have a chance I will try to learn GeoFrame. I am pround of you and your colleague. Thank you very much for share.

Since 2001, our lab has been involved in three 4D projects in China Offshore. In the past, our key point is just in seismic domain. Now, we hope to broaden horizons. our new domains are including petrophysics, seismic, well log and reservoir engineering.CNOOC is our main sponsor.

Of course, I know the importance of close-the-loop. This link(http://www.spe.org/events/dl/Huang.php) is SPE Distinguished Lecture, Dr. Huang is our collaborator. And I get some idea from his papers and lecture. This link (http://www.eage.org/...34&Opendivs=s55,s76&ActiveMenu=103) is SEG/EAGE Short Courses. Prof. Patrick Corbett also give me a lot of inspiration about integration of static and dynamic Models.

For me, a Phd student in Geophysics, my current duty just is develope a nonlinear algorithm about gird model to synthetic attributes. This algorithm is not depend on Gassmann equation and seismic wavelet. Maybe this is wrong way, but I will try.

Frankly, it is an exploration.

wzg

PS: Happy Christmas !

[ytlee]

引用框（wzgme ＠ 12-24-2010 - 07:28）

...Since 2001, our lab has been involved in three 4D projects in China Offshore. In the past, our key point is just in seismic domain. Now, we hope to broaden horizons. our new domains are including petrophysics, seismic, well log and reservoir engineering.CNOOC is our main sponsor.

Of course, I know the importance of close-the-loop. This link(http://www.spe.org/events/dl/Huang.php) is SPE Distinguished Lecture, Dr. Huang is our collaborator. And I get some idea from his papers and lecture. This link (http://www.eage.org/...34&Opendivs=s55,s76&ActiveMenu=103) is SEG/EAGE Short Courses. Prof. Patrick Corbett also give me a lot of inspiration about integration of static and dynamic Models.

For me, a Phd student in Geophysics, my current duty just is develope a nonlinear algorithm about gird model to synthetic attributes. This algorithm is not depend on Gassmann equation and seismic wavelet. Maybe this is wrong way, but I will try.

Frankly, it is an exploration.

wzg

PS: Happy Christmas !

Hello WZG,

Thanks for the link. I did a little googling on SunRise and went to the company website. I agree with Dr. Huang that we have to close the loop.

If I got the impression correct from the internet. Dr. Huang started as a geophysics student, studied reservoir engineering and worked for a geophysics company , then started his company.

I also got the impression that one of the company product is "seismic to grid" builder. If you allow me, I like to offer my past experience regarding the Re and GG in grid building. As you may already know, the major issues of computer generated grid are

1. the grid does not match the GG interpretation. ie, the grid does not look the same as the structure maps the GG staff created. The major conflicts are faults and horizon(layer) offsets. ie Y faults and reversed faults.
2. if the grid made exactly matching the paper structure map, they looks like layer cake.


Good GGRe are strong minded people. (They don't get to where they are if they don't believe in themself) Re wants the grid to fit the simulator and grids created by GG very often are not simulator friendly. On the other hand, RE grid that follows no depositional concept . The different requirement is the reason why the loop is not closing.

Right before I joined big blue, I worked on my own at Swift and got another crazy idea for closing the loop. My workflow started out with GG structure map for a layer cake grid and end with a perfect solid model that honors errorsion/deposition. My GG colleagues were floored when they saw the result thru a 3D glasses. (Now you can use a 3D printer to make a plastic model using my workflow) Uploaded is the demo version I made in 1999 to pitch my idea to public. In the zip, it contains a movie for showing a SWIFT prospect in 1998 NAPE. AFAIK, the making of layer cake grid from 2D maps have been copied by few commercial packages but none get the errosion workflow.

附加文件

•  demo.zip (3.15MB)
  下载次数: 11

[ytlee]

引用框（wzgme ＠ 12-24-2010 - 07:28）

As a new guy in Geophysics, I am sorry that I never use GeoQuest. I have watched your movie many times, and I already have a intuitive impression. If I have a chance I will try to learn GeoFrame.

....

Hello WZGME,

I dug up my GeoViz mop-up. This mop-up only read a small 3D cube. All functions are not functional except Display, and only Pan & Pick are working. Pick is the function I used to create my "Gas Chamber" movie. You can drag the left mouse to rotate and right mouse to zoom.

The original framework to read SEGY and basic display was written by my GeoQuest colleague. I made PICK work in PC and implement different PAN functions to give user better layer/slice control. To use the pick, when the dialog pops up, move the mouse to the scale, hold down the left button and draw a curve on the scale. then click on apply then rotate.

附加文件

•  MINIGV.zip (1.18MB)
  下载次数: 10

[wzgme]

引用框（ytlee ＠ 12-25-2010 - 01:59）

引用框（wzgme ＠ 12-24-2010 - 07:28）

As a new guy in Geophysics, I am sorry that I never use GeoQuest. I have watched your movie many times, and I already have a intuitive impression. If I have a chance I will try to learn GeoFrame.

....

Hello WZGME,

I dug up my GeoViz mop-up. This mop-up only read a small 3D cube. All functions are not functional except Display, and only Pan & Pick are working. Pick is the function I used to create my "Gas Chamber" movie. You can drag the left mouse to rotate and right mouse to zoom.

The original framework to read SEGY and basic display was written by my GeoQuest colleague. I made PICK work in PC and implement different PAN functions to give user better layer/slice control. To use the pick, when the dialog pops up, move the mouse to the scale, hold down the left button and draw a curve on the scale. then click on apply then rotate.

Thank you for your share. In the past three years, I also developed a simple program like this. In this post(http://forum.petro-c...92%8copendtect/), I described my UI. Recently, I will try to upload my simple code.

Thank you again!

[wzgme]

引用框（ytlee ＠ 12-24-2010 - 23:38）

引用框（wzgme ＠ 12-24-2010 - 07:28）

...Since 2001, our lab has been involved in three 4D projects in China Offshore. In the past, our key point is just in seismic domain. Now, we hope to broaden horizons. our new domains are including petrophysics, seismic, well log and reservoir engineering.CNOOC is our main sponsor.

Of course, I know the importance of close-the-loop. This link(http://www.spe.org/events/dl/Huang.php) is SPE Distinguished Lecture, Dr. Huang is our collaborator. And I get some idea from his papers and lecture. This link (http://www.eage.org/...34&Opendivs=s55,s76&ActiveMenu=103) is SEG/EAGE Short Courses. Prof. Patrick Corbett also give me a lot of inspiration about integration of static and dynamic Models.

For me, a Phd student in Geophysics, my current duty just is develope a nonlinear algorithm about gird model to synthetic attributes. This algorithm is not depend on Gassmann equation and seismic wavelet. Maybe this is wrong way, but I will try.

Frankly, it is an exploration.

wzg

PS: Happy Christmas !

Hello WZG,

Thanks for the link. I did a little googling on SunRise and went to the company website. I agree with Dr. Huang that we have to close the loop.

If I got the impression correct from the internet. Dr. Huang started as a geophysics student, studied reservoir engineering and worked for a geophysics company , then started his company.

I also got the impression that one of the company product is "seismic to grid" builder. If you allow me, I like to offer my past experience regarding the Re and GG in grid building. As you may already know, the major issues of computer generated grid are

1. the grid does not match the GG interpretation. ie, the grid does not look the same as the structure maps the GG staff created. The major conflicts are faults and horizon(layer) offsets. ie Y faults and reversed faults.
2. if the grid made exactly matching the paper structure map, they looks like layer cake.


Good GGRe are strong minded people. (They don't get to where they are if they don't believe in themself) Re wants the grid to fit the simulator and grids created by GG very often are not simulator friendly. On the other hand, RE grid that follows no depositional concept . The different requirement is the reason why the loop is not closing.

Right before I joined big blue, I worked on my own at Swift and got another crazy idea for closing the loop. My workflow started out with GG structure map for a layer cake grid and end with a perfect solid model that honors errorsion/deposition. My GG colleagues were floored when they saw the result thru a 3D glasses. (Now you can use a 3D printer to make a plastic model using my workflow) Uploaded is the demo version I made in 1999 to pitch my idea to public. In the zip, it contains a movie for showing a SWIFT prospect in 1998 NAPE. AFAIK, the making of layer cake grid from 2D maps have been copied by few commercial packages but none get the errosion workflow.

Now Dr. Huang is serving as expert of "1000 plan" for BGP.

Yes, you are right. There are many difficulties between GG and RE. From your so long work experience, I learn so much more. I very appreciate your kindly suggestion.

However, I wish I could move a small step.

[ytlee]

引用框（wzgme ＠ 12-26-2010 - 06:43）

Now Dr. Huang is serving as expert of "1000 plan" for BGP.

Yes, you are right. There are many difficulties between GG and RE. From your so long work experience, I learn so much more. I very appreciate your kindly suggestion.

However, I wish I could move a small step.

Thanks for the link. I'll definitely looking into the 1000plan. I don't know Dr. Huang, but would like to get his experience with 1000plan.

BTW, I communicated with the OPENDTECT guy who lived in Sugarland in the past. May be he knew I worked for big blue and viewed me as a foe, we did not get off on the right foot. In addition, I was more interested in another free data visualization package VTK,http://vtk.org/. Hence I have been ignoring OPENDTECT development. I look forwards to your release to catch up with OPENDTECT.

BTW, I really appreciate the discussion. I always believe the "stone soup" approach of exchanging ideas. I have been giving up software development for a long while because I have been busy with a deep water project and my interest had moved to microchip programming (http://www.microchip.com/forums/tm.aspx?m=544634&high=) . For the past few days, I have been looking over the codes and the notes I gathered over the years, my code writing fingers are itching again.

[wzgme]

引用框（ytlee ＠ 12-27-2010 - 00:24）

Thanks for the link. I'll definitely looking into the 1000plan. I don't know Dr. Huang, but would like to get his experience with 1000plan.

BTW, I communicated with the OPENDTECT guy who lived in Sugarland in the past. May be he knew I worked for big blue and viewed me as a foe, we did not get off on the right foot. In addition, I was more interested in another free data visualization package VTK,http://vtk.org/. Hence I have been ignoring OPENDTECT development. I look forwards to your release to catch up with OPENDTECT.

BTW, I really appreciate the discussion. I always believe the "stone soup" approach of exchanging ideas. I have been giving up software development for a long while because I have been busy with a deep water project and my interest had moved to microchip programming (http://www.microchip.com/forums/tm.aspx?m=544634&high=) . For the past few days, I have been looking over the codes and the notes I gathered over the years, my code writing fingers are itching again.

Yes, this news is not published in 1000 plan websites. Please click this link

引用

戴南浔博士，在中国和美国有近30年的地质与地球物理工作经验。2009年，戴南浔博士被聘为东方物探技术研究中心休斯敦研究中心副主任。加盟东方物探不到一年时间，其叠前深度偏移技术便被应用到物探生产中。
李向阳、黄旭日和梁兼栋是国务院“千人计划”引进的海外高层次人才，也是国际知名地球物理学家。3名专家加盟东方物探后，提升了东方物探在油藏地球物理和深海业务等领域的竞争力与知名度。

BTW, Maybe some forum members also remeber Dr. Dai, who was criticized in our forum because of his migration work . :D Ok, it is just a "八卦".

I very pleasure to discuss with you and other members in our forum. I very hope other one join in our discuss. I strongly think integration of GG and RE is a challenge.

For Opendtect, I just know a little as a user. I will carefully consider your question, and then response my opinion in PM box. My code is a toy that need more update. And I have no plan to maintain it.

BTW, When I was studing in middle school, I was becoming an Electronic DIYer. I have a unofficial second degree in electronic engineering . Now, I am interested in microchip platform，for example Arduino. I very want to make a intelligent automobile if I have a long free time.