LF-NMR as a tool for predicting the 3D printability of surimi-starch systems.

In this study, surimi from golden pompanos was mixed with starch to form a surimi-starch system. The water properties, rheological properties, and three-dimensional (3D) printability of the surimi-starch were measured. Cluster analysis results showed that the 3D printability was closely related to the type and addition content of starch, and the water and rheological properties. The low-field nuclear magnetic resonance (LF-NMR) parameters were used to predict 3D printability using polynomial regression models. The correlation coefficients (R2) for 3D printing accuracy and stability were 0.88 and 0.93, and the root mean square error (RMSE) values were 0.20% and 4.59%, respectively. In the verification test, the R2 for the two models were 0.85 and 0.89, and the RMSE values were 0.20% and 1.06%, respectively. The nonlinear surface regression fitting exhibited superior predictive performance. Therefore, LF-NMR is a good non-destructive tool for quickly and accurately predicting the 3D printability of the surimi-starch systems.