Prediction of textural properties of 3D-printed food using response surface methodology

Abstract

3D printing has enabled modifying internal structures of the food affecting textural properties, but predicting desired texture remains challenging. To overcome this challenge, the use of response surface methodology (RSM) was demonstrated to develop empirical models relating 3D printing parameters to textural properties using aqueous inks containing cricket powders as a model system. Regression models were established for our key textural properties (i.e., hardness (H), adhesiveness (A), cohesiveness (C), and springiness (S)) in response to three 3D printing parameters: infill percentage (i), layer height (h), and print speed (s). Our developed model successfully predicted the 3D printing parameters to achieve the intended textural properties using a multi-objective optimization framework. The predicted limits for H, A, C, and S were 0.66–5.39 N, 0.01–12.43 mJ, 0.01–1.05, and 0–19.20 mm, respectively. To validate our models, we simulated the texture of other food using our model ink and achieved high accuracy for H (99%), C (82%), and S (87%). This work highlights a simple way to 3D-print foods with spatially different textures and materials, unlocking the full potential of 3D printing technology for manufacturing a range of customized foods.