Predicting hydrogen production in porous foams for steam methane reforming: A combined approach using computational fluid dynamics and machine learning regression models