leveraging machine learning to improve ... - geo-expo.comthe information contained in this...

Confidential. © 2019 IHS Markit®. All Rights Reserved.Confidential. © 2019 IHS Markit®. All Rights Reserved.

Leveraging Machine Learning To Improve Geological and Petrophysical Workflows

Malleswar Yenugu, PhDAssociate Director in Data AnalyticsIHS Markit, Houston, USA

Confidential. © 2019 IHS Markit®. All Rights Reserved. 2

AI and ML 2

Machine Learning

~

Confidential. © 2019 IHS Markit®. All Rights Reserved.

AI for Upstream – E&P 3

3


Outline 4

4

➢ The Challenges for Geoscientists

➢ Log Predictor for Missing Log Generation

➢ Geology based ML Log Prediction

➢ Auto Top Picker for Formation Tops

➢ Auto Facies Predictor

➢ Summary


The Challenges for Geoscientists

5


The Challenges: Missing Logs or No Logs 6

6

??

6


The Challenges: Basin wide Formation Tops Picking 7

7


The Challenges: Facies Identification 8

8

Neutron Porosity (v/v) PEF (Barns/Electron)

DT

(us/

ft)

RH

OB

(g/c

c)


Log PredictorTM using ML

9


Linear Regression and Random Forest ML Algorithms 10

. . ..

. .. .. .. . .. . .. ... . . .

..

... .. .. . .. . .

. . .. .. .. . ..

.. .. ...

. . ..... ..

. ... .

...... .... . .. .. ..... ... . ..... . ..

.. .. . .. ... . .

.. .. . .. . .. ... .. . ..

. ... .

Logs used in the Training Log to Predict Linear Regression Predicted Log

Logs used in the Training Log to Predict Random Forest Predicted Log

10


XGBoost Supervised Learning Algorithm 11

XGBoost…..

Errors Errors

11


Log PredictorTM : Powder River Basin

12


Leveraging ML: Powder River Basin 13

➢ A total of 273 wells used in the ML training

➢ Few key logs are missing in the well selected for ML prediction

➢ Core data is also used in the ML training

13


Log Prediction – Comparison of ML Algorithms 14

14


15

Derived from DeltalogR technique

TOC Log Prediction using ML

15


Missing Log Generation using ML 16

16


Actual vs. ML Predicted Log 17

TOC (wt%)

ML

Pre

dict

ed T

OC

(wt%

)

ILD (Ohm-m)

ML

Pre

dict

ed IL

D (O

hm-m

)17


Log PredictorTM : Powder River Basin –Reservoir (Geology) Based Prediction

18


What Logs We want to use for ML Training and Prediction?

19Geology (Reservoir/Physics) Based Log Prediction

19


20Geology Based Log Prediction

20


21

Actual vs. ML Predicted (based on Geology) Log

ILD (Ohm-m)

ML

Pre

dict

ed (b

ased

on

Geo

logy

) ILD

(Ohm

-m)

21


Guideline Chart for Geology based Log Prediction 22

22


Auto Top PickerTM using ML

23


24Autotop PickerTM - CNN

CNN

24


Auto Top PickerTM – Appalachian Basin

25


26Auto Top PickerTM - Marcellus

GR: 1624 wells

GR&RHOB: 667 wells

Total No. of wells used in the Training (picked manually): 164

26


27Tops from a Geologist

W1

W2W3

W4

W5 W6 W7

27


28ML picked Tops – Auto Tops PickerTM

W1

W2

W3

W4W5

W6

28


29Human-ML Tops Picks Comparison

Human picked Tops

Machine picked Tops 29


Auto Facies using ML – Work in Progress

30


31Permian basin well

BS

_LM

BS

1_S

DB

S2_

SD

BS3_

SDW

olfc

amp

31


32Auto Facies using ML

DT (us/ft) DT (us/ft)

GR

(AP

I)

GR

(AP

I)

Kmeans Unsupervised ML Clustering

32

1

2

3

4

5


33Auto Facies using ML

NPHI (v/v)R

HO

B (g

/cc)

RH

OB

(g/c

c)

Kmeans Unsupervised ML Clustering

33

1

2

3

4

5

NPHI (v/v)


Summary

34


Summary 35

➢ML is an additional tool to help Geoscientists

➢Both ML and Geology based Approach

➢Supervision from Geoscientists is necessary to reduce the Uncertainty

35

IHS Markit Customer [email protected]: +1 800 IHS CARE (+1 800 447 2273)Europe, Middle East, and Africa: +44 (0) 1344 328 300Asia and the Pacific Rim: +604 291 3600

Disclaimer

The information contained in this presentation is confidential. Any unauthorized use, disclosure, reproduction, or dissemination, in full or in part, in any media or by any means, without the prior written permission of IHS Markit Ltd. or any of its affiliates ("IHS Markit") is strictly prohibited. IHS Markit owns all IHS Markit logos and trade names contained in this presentation that are subject to license. Opinions, statements, estimates, and projections in this presentation (including other media) are solely those of the individual author(s) at the time of writing and do not necessarily reflect the opinions of IHS Markit. Neither IHS Markit nor the author(s) has any obligation to update this presentation in the event that any content, opinion, statement, estimate, or projection (collectively, "information") changes or subsequently becomes inaccurate. IHS Markit makes no warranty, expressed or implied, as to the accuracy, completeness, or timeliness of any information in this presentation, and shall not in any way be liable to any recipient for any inaccuracies or omissions. Without limiting the foregoing, IHS Markit shall have no liability whatsoever to any recipient, whether in contract, in tort (including negligence), under warranty, under statute or otherwise, in respect of any loss or damage suffered by any recipient as a result of or in connection with any information provided, or any course of action determined, by it or any third party, whether or not based on any information provided. The inclusion of a link to an external website by IHS Markit should not be understood to be an endorsement of that website or the site's owners (or their products/services). IHS Markit is not responsible for either the content or output of external websites. Copyright © 2019, IHS Markit®. All rights reserved and all intellectual property rights are retained by IHS Markit.

36

Leveraging Machine Learning To Improve Geological and Petrophysical Workflows

Malleswar Yenugu, PhD

leveraging machine learning to improve ... - geo-expo.comthe information contained in this...

Documents