Data-driven modeling and monitoring of fuel cell performance

A mathematical framework that provides practical guidelines for user adoption is proposed fuel cell performance evaluation. By leveraging the framework, two measures describe average and worst-case are presented. To facilitate computation of in a setting, we model distribution voltages at different current points as Gaussian process. Then minimum number samples needed to estimate obtained using information-theoretic notions. Furthermore, introduce sensing algorithm finds maximally informative about voltage. Observing identified by enables unobserved points. The corresponding results validated on dataset provided an industrial whose conclusion coincides with judgement from manufacturer.