Development of a rheological prediction model for food suspensions and emulsions

► The rheological behavior of 34 commercial food dispersion is modeled. ► Artificial neuronal networks are used to successfully predict rheological properties. ► Increasing product database will improve predictive capabilities. The rheological behavior of 34 commercial food dispersions was investigated and modeled with the Herschel–Bulkley model. Artificial neuronal networks (ANNs) were trained to predict the rheological parameters yield stress τ0, consistency coefficient K and flow behavior index n in dependency of the composition of fats, carbohydrates, proteins and water. ANNs with 3 hidden layers and 2 neurons per layer showed good to very good results for all Herschel–Bulkley parameters.