IBM and FuelCell Energy, Inc. will work together to boost the performance of FuelCell Energy’s technology using Foundation Models, a form of generative Artificial Intelligence (AI). The aim of this collaboration is to support both companies’ efforts to lead a global transition to renewable energy sources that emit little to no carbon.
Foundation models, a term popularized by the Stanford Institute for Human-Centered Artificial Intelligence, are intended to replace the task-specific models that dominate AI to date. Foundation models can be trained on a broad set of unlabeled data that can be used for different tasks, with minimal fine-tuning.
Through the collaboration, IBM will research ways that FuelCell Energy can extend the life of its fuel cells through an optimal control of operational parameters and their cost effectiveness for customers.
IBM will develop a Foundation Model to create device-level models using FuelCell Energy data. These models will learn from the data and predict the technology’s performance. This creates a data-based digital twin of the fuel cell, connecting various data sources and helping to provide a greater understanding of how various operating parameters impact the fuel cell’s degradation, enabling potential improvements.
FuelCell Energy’s Chief Technology Officer, Tony Leo, and Alessandro Curioni, VP IBM Research Europe and Africa and Director of IBM Research Europe, Zurich, announced the collaboration at the 2023 Pan-EMEA Media Event at IBM Research Zurich lab in Switzerland.
We hope that IBM’s AI will help FuelCell Energy replace the traditional, time-intensive and expensive accelerated life testing process when it comes to electrochemical energy production to quickly and efficiently propel forward the world’s transition to clean energy.—Alessandro Curioni
FuelCell Energy holds more than 450 fuel cell technology patents in the United States and globally.