Applied AI in Energy



Climate change and depletion of the Earth’s natural resources are frequent media headlines directly tied to demand for clean, affordable, and reliable energy. In parallel, gains in computing, memory, and storage have made artificial intelligence technologies more accessible. The intersection of AI and energy may hold the key to unlocking some of the greatest challenges our world faces. “Exponential technology is rapidly pushing electricity to reach the point of eventually becoming nearly free,” remarks Pascal Finette of Singularity University. Research into renewables and the use of simulations to creates digital twins are two examples of the accelerated value that machine learning brings to the energy sector. This track explores the tremendous promise possible today and in the near future.

Thursday, October 24

7:45 am Registration Opens

8:00 Continental Breakfast

8:50 – 12:15 pm Keynote Session

12:15 pm Networking, Coffee & Dessert in the Expo

1:30 pm Opening Remarks

Prouty_KevinKevin Prouty, Group Vice President IDC Energy Insights, IDC


1:35 AI Executive to be Announced

2:05 Beyond Asset Automation: AI as a Transformational Foundation in the Energy Industry

Prouty_KevinKevin Prouty, Group Vice President IDC Energy Insights, IDC

This session highlights how AI and process automation plays a key role in transforming the energy industry. From real-time decision-making at the well site in the Oil & Gas industry to automated outage management in utilities, AI is in the process of remaking how people interact with their businesses. Use cases that will be explored range from self-healing production assets to automating truck rolls for field service, to revenue protection with smart meters. Learn how companies throughout the energy value chain are using AI to solve problems long considered unsolvable.

2:30 Using AI to Improve Industrial Energy Efficiency

Piva_DaianeDaiane Piva, Energy Efficiency Improvement Consultant, Tata Steel Europe

The fact that AI is set to disrupt industrial sectors is no news, but the specific ways it will do so is still left for discovery. With that in mind, our team has taken on the task of exploring the power of AI to improve the energy performance of industrial processes. What we have found was beyond expectations. Using millions of data points from re-heating furnaces, we identified the most important parameters that triggered energy use, and based on that, trained an AI model which optimized consumption to drive costs and CO2 emissions down.

2:55 Addressing Multi-Agent Challenges in Energy Systems with AI

Franklin_MichaelMike Franklin, PhD, Assistant Professor, College of Computing, Kennesaw State University, Marietta Campus

The harsh reality is this - energy system modeling is highly-complex. Many AI systems use monolithic models and uniform policies, both of which reduce the efficacy and applicability of the system. Multi-agent systems, where each agent or aggregate group of agents have their own distinct policy, are more powerful and expressive, and thus model the complex world of energy systems more effectively. Further, they can better reflect the hierarchical nature of these systems. 

3:20 Networking Break in the Expo

4:05 Transforming Asset Inspection in Energy leveraging Computer Vision & NLP

Krishnaswamy_ShyamShyam Krishnaswamy, Director of Innovation & Strategy, Exelon Corporation 

Asset performance is key to power industry operations. Optimizing and maximizing the life of assets become very important in controlling costs, delivering service to our customers. Aerolabs, an Exelon company, is leveraging drones and AI capabilities inspecting and detecting anomalies to drive asset performance. 

4:35 Using Physical Sensor Data at Scale for Informed Decision-Making

Zeroug_SmaineSmaine Zeroug, PhD, Research Director, Applied Math and Data Analytics, Schlumberger

While progress has been made in the application of ML to measurement data pertaining to subsurface characterization in O&G exploration and field development, this has largely focused on bringing efficiency to time-consuming processing tasks. Application of deep learning and AI to these data for developing cognitive interpretation algorithms faces multiple challenges due to the data scarcity, richness in information, and lack of labels. Domain knowledge and physical modeling of subsurface measurements and geological processes become critical ingredients to address these challenges.

5:05 Session Break

5:35 Networking Reception in the Expo

6:35 Meetup Groups

7:35 Close of Day 2