Machine learning enables more efficient (nano)catalyst management, enhancing the competitiveness of (bio)hydrogen production from sewage sludge

Dark fermentation is touted as an environmentally sustainable method of (bio)hydrogen production, however, its economic sustainability is questionable. The main reason is the problematic stability and low yield of the process accompanied by the formation of byproducts with inhibitory effects. The current study investigates the potential of nickel oxide (NiO), iron nanoparticles (Fe2O3) and their combination with activated carbon to enhance hydrogen production in a dark fermentation system using anaerobic sludge. The experimental setup includes the 14 scenarios that analyze the interactions of promising (nano)catalysts using free and immobilized cell technology Various parameters such as cumulative hydrogen production, daily hydrogen production, hydrogen yield, oxidation-reduction potential, and post-fermentation pH were analyzed. The Modified Gompertz model was used to evaluate the kinetic parameters of the dark fermentation process. In addition to above, random forest machine learning model was applied to predict the production of hydrogen. The results demonstrated that the supplementation of nanomaterials, significantly enhanced hydrogen production compared to the control creating better prerequisites sustainable hydrogen economy.