U.S. Department of Energy Adds One More Team to Advanced Machine Learning Effort for Solar

Stony Brook University, Pacific Northwest National Laboratory and Ecogy Energy O&M focus highlights the importance of driving down real-world soft costs

DOE, Ecogy Energy, Stony Brook, PNNL

Ecogy Energy "Ecogy," a Brooklyn-based, vertically-integrated Distributed Energy Resource ("DER") owner/operator, announced that it is part of a research team that was selected to receive $230,000 of a $750,000 award from the U.S. Department of Energy Solar Energy Technologies Office ("SETO") to advance artificial intelligence ("AI") research and development for solar energy applications. This project will apply machine learning technology to commercial & industrial ("C&I") portfolios in order to reduce O&M costs. 

Commercial industrial solar portfolios represent over 20% of the industry's total fleet, with the entire industry forecast to triple over the next five years. In order to operate these assets effectively, machine learning ("ML") advances are promising, but often out of reach for multi-vendor footprints due to complexity and cost. Ecogy, Stony Brook and Pacific Northwest National Laboratory ("PNNL") aim to solve this problem using an open-source platform that boosts interoperability both for devices from multiple manufacturers and multiple ML frameworks. 

"We are hyper-focused on getting our C&I portfolios operating at top performance," says CEO Jack Bertuzzi, who operates Ecogy's DER portfolios, which are spread across the East Coast, from the top to the bottom, including the US Virgin Islands. "We began by opening up communications for monitoring and control of all our assets, normalizing the dataset, and finally, we upped the resolution. Now with our partners Stony Brook and PNNL, we're going to forensically comb all of that data to create actionable information and extract tremendous value through ever-expanding insights."

Ecogy's machine learning solution is based on the SolarNetwork platform, an open-source software platform highlighting interoperability and reduced costs for managing systems with complex topologies. Ecogy, Stony Brook and PNNL aim to provide the whole industry with a toolkit to help them operate their assets more effectively; all of these developments will be released as open source to maximize access to these results.

"Our first goal is to deliver advanced ML techniques for solar plants so asset managers can respond quickly and efficiently with new insights about their assets, delivered in real time or in some cases before noticeable production-impacting events begin to occur. The second goal is to simplify the deployment of such software so it can be widely implemented without the need for ML expertise. What's really great is that we're using an incredible live real-world data set to continually test our hypotheses," says Yue Zhao, project leader at Stony Brook.

Ecogy was selected as a part of the SETO Fiscal Year 2020 funding program, an effort to advance research and development projects that will lower solar electricity costs, increase the competitiveness of American solar manufacturing and businesses, improve the reliability and resilience of the grid, and expand solar to new applications. Ecogy's is one of several projects that will leverage U.S. AI experience, especially in the area of machine learning, to develop disruptive solutions across the solar industry value chain by forming partnerships between AI experts and industry stakeholders.

Ecogy, Stony Brook and PNNL aim to drive down the "soft costs" of deploying and managing C&I solar assets to make them a highly relevant asset class in the renewables industry. 

About Ecogy Energy
Ecogy was founded as a developer, financier and owner-operator of distributed generation resources in 2010. Ecogy have since found their niche specializing in financing distributed energy assets for traditionally underserved entities including affordable housing, non-profits and municipalities.


Source: Ecogy Energy