Artificial Neural Networks to Optimize Oil-in-Water Emulsion Stability with Orange By-Products

Abstract

The use of artificial neural networks (ANNs) is proposed to optimize the formulation of stable oil-in-water emulsions (oil 6% w/w) with a flour made from orange by-products (OBF), rich in pectins (21 g/100 g fresh matter), in different concentrations (0.95, 2.38, and 3.40% w/w), combined with or without soy proteins (0.3 and 0.6% w/w). Emulsions containing OBF were stable against coalescence and flocculation (with 2.4 and 3.4% OBF) and creaming (3.4% OBF) for 24 h; the droplets’ diameter decreased up to 44% and the viscosity increased up to 37% with higher concentrations of OBF. With the protein addition, the droplets’ diameter decreased by up to 70%, and flocculation increased. Compared with emulsions produced with purified citrus pectins (0.2 and 0.5% w/w), OBF emulsions exhibited up to 32% lower viscosities, 129% larger droplets, and 45% smaller Z potential values. Optimization solved with ANNs minimizing the droplet size and the emulsion instability resulted in OBF and protein concentrations of 3.16 and 0.14%, respectively. The experimental characteristics of the optimum emulsion closely matched those predicted by ANNs demonstrating the usefulness of the proposed method.