Tailoring the Properties of Pea-Enriched Soft Cakes Using a Multiobjective Model Based on Sensory-Relevant Instrumental Characterization

Managing the quality of pea-enriched cakes made from mixes of wheat and pea flours in various proportions (10, 35, and 60 wt% of pea, flour basis) and with various particle size distributions (0, 50, and 100 wt% of particles < 63 µm) is a challenge for the industry. A “multiobjective” model based on an I-optimal response surface design was set up in a previous study. It allows obtaining target cake structural and textural properties by adjusting several processing parameters (mixing speed and time, baking program). As the model’s ability to correct the variations in cake properties due to variations in flour properties remained to be proven, two case studies concerning the proportion and the particle size of pea flour were studied. A variation of crumb stiffness (24 to 37 kPa), lightness (85.4 to 79.6 in L*), and cell fineness (4.9 to –4.8 in PC1 score) could be observed with the increase in the proportion of pea flour from 0 to 35 wt%, and these variations were properly corrected by the model (corrected values: 28 kPa; 83.2 in L*; 3.7 in PC1 score). A change in the particle size of pea flour caused variations in cake properties inferior to those due to processing reproducibility, except for cake symmetry (7.5 to 10.1 in symmetry index; corrected value: 7.2). A selection of products representative of the diversity of cakes from the original design space was investigated by 11 trained panelists through quantitative descriptive analysis. A convergence between sensory and instrumental results was found concerning structural and textural properties. Additional sensory perceptions such as beany attributes or in-mouth drying aftertaste were pointed out.