Machine Learning, Deep Learning & AI in Oil and Gas - From the Energy Conference Network

This event ran April 19-20, 2016 and has now finished.

If you have a case study that you would like to present, please contact:

Melissa Zerber | Conference Producer
Energy Conference Network
+1 (855) 869-4260

7:30 am

Registration and coffee

8:30 am

Chairperson opening remarks

Mark Reynolds, Senior Solutions Architect Technical E&P Applications
Southwestern Energy

Innovation and Implementation

8:45 am

Machine Learning: “You want the truth? You can’t handle the truth!”

Machine learning (ML) has had a slower adoption rate in the oil and gas industry and though there are many supporters there are also many skeptics about the real value it can bring. This presentation will show proof of concepts, applications in the field, challenges and solutions for ML within the oil and gas industry, and will also seek to answer many of the common questions raised. What is the reality of ML for companies in the oil and gas industry? Are we on the hype cycle? We heard that Google was doing it so we should try it as well? Do our engineers and scientists believe that a machine can solve the problems we encounter accurately enough to consider using ML? Oil and gas sits at Andy Grove, Co-Founder of Intel’s Inflexion Point – Is ML going to drive the us to new heights or …? What areas are going to be the first adopters of ML? What are some of the examples in the field that are currently being used?

Sammy Haroon, Director, Palo Alto Innovation Center and Director, Global Enterprise Data Analytics Group
Baker Hughes

9:15 am

Data Babble

Today’s data is often lost to those who need it, disorganized when it is needed quickly, and incomprehensible to people whose job is to understand it, or just plain wrong.  New technology promises to solve those problems, make us efficient, and make us richer.  All too often, we fail to realize the value the “silver bullet” hyped promised, as implementation

Like any edifice, technology requires a solid foundation if it’s to be successful; in our industry, the foundation is data. This talk focusses on the problems we have with data and what needs to be done about it, so that advancing technologies can achieve their potential in our industry.

Tarun Chandrasekhar, Technology Solutions Director for Business Development and Exploration
BP Lower 48 Onshore

9:45 am

Morning break and networking

10:30 am

Case Study - Running Internal Data Science Challenges - How Chevron bought together data science competitors to discover new insights and potential solutions

A competition that brings together a group of data science competitors to address a problem stands a better chance of finding the best solutions.  Chevron has run two internal data science challenges – each open only to company employees and addressing an actual business problem. The challenges were conducted in phases:

  • Startup – labeled and blinded data and documentation were made available to contestants
  • Model development – contestants developed models with labeled data and made predictions on blinded data. Predictions were scored and model accuracy feedback relative to other competitors was provided via a challenge website
  • Final evaluation – contestants selected their best models for final evaluation, and made predictions on a distinct, blinded dataset. These were scored to determine challenge winners.

Both challenges resulted in new insights and potential solutions to the problems addressed, and in the identification of analytical talent within Chevron.

Cole Harris, Data Scientist
Chevron

11:00 am

Best Practices in IIoT, IT/OT Integration, and Leveraging Machine Learning and Advanced Analytics to Deliver Business Value

Best practices in choosing, designing, and implementing from the portfolio of analytical methods including real-time, geospatial, machine learning, statistical, and “big data”

  • How can we shift our time and focus from ensuring data quality and data preparation to performing advanced analytics and delivering business value?
  • How do we operationalize the results from our advanced analytics and models and perform plan vs actual financial based analytics?
  • How can business value be rapidly attained in scale and sustained over time?
  • How to we integrate IoT/IIOT, pervasive sensing, and the cloud into my companies existing data fabric security and with context and exploit advanced analytics?
  • How to we integrate our connected ecosystem of suppliers, customers, and stakeholders and realize the vision of the “digital value chain”?

Craig Harclerode, Global O&G and Industrial Chemicals Business Development Executive
OSIsoft
Curt Hertler, Global Solutions Architect
OSIsoft
Sameer Kalwani, Founder
Element Analytics

11:40 am

Road Map to Constructing a Top Down Big Data and Machine Learning System

E&P organizations are turning more attention to accumulated data to enhance operating efficiencies, safety, and recovery. The computing paradigm is shifting, the O&G paradigm is shifting, and the rise of the machine learning paradigm requires careful attention to topdown integrated systems engineering. A system approach will be presented to stimulate out-of the-box thinking to address the machine learning paradigm.

Mark Reynolds, Senior Solutions Architect Technical E&P Applications
Southwestern Energy

12:10 pm

Networking Lunch

1:45 pm

Industrializing Machine Learning as a key tool in the midst of Chaos

Kenneth Smith, General Manager, Oil & Gas
Hortonworks
Dr. Arvind Battula, Sr. Data Scientist
Schlumberger

2:15 pm

Disrupting a Legacy Industry

  • Disruptive technologies poised to reinvent the oil industry
  • Deep Learning and Machine

Nav Dhunay, CEO
Ambyint

2:45 pm

What it takes to create value from Machine Learning for E&P?

Machine learning has been around for few decades and E&P industry has used various algorithms for a diverse set of problems. However, not a significant amount of industry problems have been solved effectively.  The challenges lies in how we leverage machine learning.

E&P industry needs to leverage an integrated and modular approach to address industry problems. They need to bring search, automation, to build recommendation for actions from insights from machine learning applied to E&P Big Data. Uses case related drilling performance, and real-time big data analytics to get insights into NPT and ILT will be briefly discussed.

Dr. Satyam Priyadarshy, Chief Data Scientist
Halliburton

3:15 pm

Drinks reception sponsored by Hortonworks

8:00 am

Morning coffee and registration

9:00 am

Chairperson opening remarks

Mark Reynolds, Senior Solutions Architect Technical E&P Applications
Southwestern Energy

9:10 am

How Machine Learning Complements and Enables the Industrial Internet of Things

Steve Jennis, Corporate VP and Head of Global Marketing
ADLINK Technology & contributor for the Industrial Internet Consortium

Optimization

9:40 am

Automating Real-Time Adaptive Edge Analytics

Learn how ThingWorx, Flowserve, National Instruments and HP Enterprise came together to create a powerful piece of technology that continuously monitors, adapts and operationalizes several different types of complex analytical techniques that all oil and gas companies can benefit from.  In this session, you will see real-time edge analysis on data streams, predictive intelligence and simulations that enable engineers and operators to better create, service and operate their machinery in their environments – all without any manual coding of advanced analytical techniques or machine learning processes.

Eric J. van Gemeren, Vice President, R&D
Flowserve Corporation
Eric Smith, Vice President of Business Development, Technology Platform group-Machine Learning
Thingworx-PTC

10:10 am

Utilization of Machine Learning in New Production System Selection

  • Machine Learning utilization to analyze field data and forecast production system performance
  • Benefits of historical data utilization in Artificial Lift and Chemicals businesses

Emanuel Marsis, Production Modeling & Simulations Engineer
Baker Hughes

10:40 am

Morning networking and break

11:10 am

Improve Forecasting Asset Performance and Asset Health through Implementing Predictive Models

  • Building models that use historical data to predict asset performance
  • Continuously improving models using historical data

Stuart Gillen, Director of Business Development
Spark Cognition

11:40 am

Innovation Showcase

Dr. Eric Schoen, Director of Engineering
i2k Connect
Ray Richardson, CTO
Simularity
Carl Byers, Co-Founder and Chief Strategy Officer
Contextere

12:40 pm

Networking Lunch

1:40 pm

Enhancing Equipment Monitoring Capabilities Using Anomaly Detection

  • Processing data in real time to improve anomaly detection
  • Ensuring algorithms are up to the challenge

Subrat Nanda, Senior Data Scientist & Analytics Leader
GE
Matthew Krueger, Analytics Engineering Leader for Distributed Power
GE

2:10 pm

Scaling Models to Work Across Global Assets

Opportunities for translating established Machine Learning models from one asset to another. Consideration of similarities and differences across global asset or fleet types; the need to modify predictions based on unique vs generic process designs

Dr. Gilbert Haddad, TLM Analytics Manager
Schlumberger

2:40 pm

Data driven solutions to reduce unplanned downtime

  • Obtaining real time insight, contextualizing and visualizing the data to eliminate unexpected failures
  • Applying historical data and using effective modeling to connect data points previously not evaluated to  predict asset performance and prevent failures

Usman Shuja, Vice President of Market Development
SparkCognition
Eric J. van Gemeren, Vice President, R&D
Flowserve Corporation

3:40 pm

Closing remarks and end of conference