Optical determination of the biomass particle size for Cellulosic Biofuel Production

Cellulose-based biofuels are currently the most promising solution for replacing petroleum-based liquid fuels, which are mainly used in the transport sector. In view of the fact that petroleum-based fuels are a finite resource, the search for alternatives and the efficient design of manufacturing processes for these is of considerable importance. The particle size of the biomass is a decisive parameter in the production of cellulose biofuels and, depending on the type of biomass, can have a significant influence on the sugar yield during hydrolysis. The resulting biofuel yield after the fermentation process correlates almost proportionally with the sugar yield of the hydrolysis. The high number of particles in a biomass sample poses a challenge for the fast and precise determination of the average particle size. This study investigates whether the particle size of a biomass sample can be reliably determined using Convolutional Neural Networks (CNNs). The results indicate that optical detection of the particle size of maize-based biomass samples is feasible in principle.