My work in the field of Artificial Intelligence - Episode 2 (1998-2004)
I joined SLB in 1998. SLB is biggest oilfield services company in world. Obviously I could have hoped to get best technology exposure in that company. And I always kept looking for any application that SLB introduces using Artificial Intelligence.
SeisClass was first such application introduced by SLB, that used artificial intelligence, as one of the module under seismic interpretation package Charisma within its flagship geoscientific platform Geoframe. All these offerings are now retired.
In 2000, SLB got first seismic interpretation and reservoir characterization project in India from Selan Exploration Technologies LTD for their Bakrol Field in India. Below is field location in Map
At that time the field only had 7 wells, 5sqkm of 3D seismic data around most producing well, some 2000LKM 2D seismic data.
It is this project where I used SeisClass to predict and map potential good reservoir not only in 3D coverage but 2D coverage too using the information within overlap area. And it did show large pool of good reservoir in area covered by 2D seismic. My approach was largely criticized at that time by seismic experts within SLB as well as one of SELAN expert. It was very difficult, for relatively inexperienced, to champion the use AI based method for a commercial project. But I found great support in my immediate manager who had supported me throughout. Although, I now believe that some of the criticism was valid but under those limitations that was the best approach. I am not privy to the later development in the field but from public information from Selan Website when I see that
"As on March 2017, there are 23 producing wells in the Bakrol Field and daily total production is about 600 barrels and 20,000 SCM of oil and gas per day respectively. Bakrol Field has produced nearly 2.09 MMbbls of crude oil and 73 MMscm of associated natural gas cumulatively after commencement of operations post taking over the block from ONGC in October 1995"
I feel confident that many of the producing wells must lie within the predicted good quality area at that time covered by 2D seismic only. Moreover, the study boasted the confidence of the Selan owners at that time to invest in further E&P activity in the field for which they are rewarded handsomely.
For me - I used this work as my GFE (General Field Engineer) project. During the presentations of the project at local level and at area level I again faced many critics of the approach but at last It was approved and I got my GFE promotion from Grade 10 to Grade 11 in Schlumberger in 2002.
One important question that I was not able to answer during internal presentation was from Nader C. Dutta . He asked me - where is the low frequency model in my entire analysis. And to be honest I had no clue what was the question. I now know the importance of low frequency model (LFM) in seismic interpretation and reservoir characterization projects. I will discuss that in detail in some other post. However, I now also can say that LFM is not relevant when we are working on attributes computed on a seismic horizon and analysis is restricted on that horizon. So the results obtained were correct.
One more module that came up under Geoframe using Artificial Intelligence was RockCell. I tried to use that during my Indian Assignment but did not find much support from Clients and management. I found the opportunity to use it and some more during my Cairo assignment where I was transferred in 2004. I will discuss that in next post.
During this period I also published two technical paper on how to use artificial intelligence for reservoir characterization.
- Tyagi P., Bhaduri A.: Artificial Neural Networks - A tool to integrate exploration and production data for reservoir description (2003), Petrotech'2003,New Delhi, India"
- Tyagi P., Dutta D., Bhaduri A.,: Predicting permeable zones in carbonate reservoir from seismic : Addressing the challenge (2004), SPG, Hyderabad, India
A post was published for second paper in below link.
Predicting Permeable Zones In Carbonate Reservoir From Seismic - Addressing The Challenge
Keep steeming.
Previous Posts on this story
My work in The field of Artificial Intelligence - Episode 1 (1990-1995)
Note
Any relevant comment will be upvoted.
Things have changed since I worked with this stuff in GISc
Posted using Partiko Android
Yes but basic concept still remain same i.e. supervised learning and unsupervised learning.
This post has been voted on by the SteemSTEM curation team and voting trail.
If you appreciate the work we are doing then consider voting us for witness by selecting stem.witness!
For additional information please join us on the SteemSTEM discord and to get to know the rest of the community!
Great post!
Thanks for tasting the eden!
You got a 38.45% upvote from @redlambo courtesy of @prameshtyagi! Make sure to use tag #redlambo to be considered for the curation post!
You got a 40.85% upvote from @emperorofnaps courtesy of @prameshtyagi!
Want to promote your posts too? Send 0.05+ SBD or STEEM to @emperorofnaps to receive a share of a full upvote every 2.4 hours...Then go relax and take a nap!
You have been defended with a 69.19% upvote!
I was summoned by @prameshtyagi.
This post has received a $54.55 % upvote from @siditech thanks to: @prameshtyagi.
Here's a banana!
This post has received a 15.65 % upvote from @boomerang.
This post has received a 45.50% upvote from @lovejuice thanks to @prameshtyagi. They love you, so does Aggroed. Please be sure to vote for Witnesses at https://steemit.com/~witnesses.
Congratulations, your post received 43.78% up vote form @spydo courtesy of @prameshtyagi! I hope, my gratitude will help you getting more visibility.
You can also earn by making delegation. Click here to delegate to @spydo and earn 95% daily reward payout! Follow this link to know more about delegation benefits.