Petrophysics Intelligence and Automation – A Case Study of Utilizing Machine Learning to Integrate Multi-Scale Data for TOC Characterization in Unconventional Reservoirs
Westside Houston
Speaker:
Seminar Date: Aug 24 2023
Registration Opens: Jul 27 2023 - Aug 22 2023
Time: 11:30 AM - 01:00 PM (US CDT)
Admission/Registration Link: None
Meeting/Webinar Link: None
Contact: Neal Cameron (VP Westside, SPWLA Houston Chapter)
Corresponding: vpwestside@spwla-houston.org
Fees:For Student: $35
For Member: $35
For Non-member: $35
NOTES:
Speaker : Chicheng Xu
Date : Thursday, Aug 24, 2023
Time : 11:30 am – 1:00 pm (US CDT)
Venue : GEOLOG Americas, 10402 Valley Forge Dr., Houston, TX 77042
Admission : Please register using the link below.
Parking Info : Guest parking is available free of charge.
The seminar will not be open for online registration since it is an in person only event.
After registering, you will receive a confirmation email containing information.
In person attendance is limited to 30 people. Advanced registration is preferred and cut when reaching 30.
Please register by Aug. 23, 2023 @ 5pm to reserve lunch for those attending in person.
Contact : Neal Cameron (SPWLA Houston VP Westide)
Corresponding vpwestside@spwla-houston.org
ABSTRACT:
Petrophysics Intelligence and Automation (PIA) leverages advanced technologies, such as artificial intelligence (AI), machine learning (ML), data analytics (DA), and computer vision (CV), to automate and optimize various petrophysical characterization workflows. In this paper, we present a case study that utilizes various above-mentioned techniques to integrate multi-scale well data including core photos, core scans, core plugs and well logs to characterize total organic carbon (TOC) in unconventional reservoirs. We introduce a ML assisted automatic petrophysical workflow that converts core photos into continuous quantitative features that can be integrated with routine core analysis and well logs for integrated reservoir characterization. Rock types and their associated properties such as TOC can be predicted based on the quantitative attributes of core photos such as color, brightness, and texture variations by using ML algorithms such as k-means clustering and support vector regression. The workflow consists of multiple steps including training data optimization, automatic core-log depth shift, image attributes processing, and support vector regression. We applied the workflow to characterize unconventional reservoirs based on multi-scale well data from the Eagle Ford Shale USA including core photos, and core plug measurements of petrophysical and geochemical properties. Inclusion of quantitative, continuous, and high-resolution image attributes significantly enhanced the accuracy of both rock type classification and rock properties prediction.
BIOGRAPHY:
CHICHENG XU joined Aramco Houston Research Center in 2017 and is working as a research petrophysicist in the AI Technology Group. His research is focusing on petrophysics intelligence and automation using advanced computational techniques and data analytics for interpretation, classification, and modeling based on multi-scale subsurface data integration. He earned his PhD degree at the Petroleum & Geosystems Engineering Department of UT Austin in 2013 and worked as a petrophysicist/rock physicist for BP America and BHP Billiton from 2013 to 2017. He co-founded and chaired the SPWLA PDDA SIG and initialized a student scholarship for PDDA related graduate research. He also served as associate editor for several international scholastic journals including Petrophysics and SPE Reservoir Evaluation & Engineering. He was selected to receive the regional Formation Evaluation technical award by SPE - Gulf Coast in 2018, the SPWLA meritorious service award in 2019, the SPE outstanding associate editor Award in 2020, the SPWLA meritorious technical award in 2021, and the regional Data Science and Engineering Analytics technical award by SPE – Gulf Coast in 2022.