Estimation of milk fat globules rising up using NIR scattering

The size of the fat globules substantially influences the the milk skimming process. The content in phospho and sphingolipids of the fat of the cheese it is strongly influenced by the size of the fat globules. Parmesan cheese, for example, is obtained from partially skimmed milk by natural rising up of fat globules. Changing time and temperature of the rising up process can change not only the fat content of the cheese, but also greatly increase the phospholipids content. A mathematical model was developed in order to calculate the amount of skimmed fat using the NIR estimate of the fat globules distribution according to Cabassi 2013. The model takes into account the geometry of the skimming container, the initial concentration of the fat, the temperature and time of rising up. The model was then validated by the 20 experiments conducted in thermostated conditions, using both individual and mass milk.
The model calculates, according to Stokes' law, for each diameter class, the limit depth from the free surface from which the fat globules of each diameter class can rise up at a particular time . Particles of a diameter class below this limit depth will remain in suspension after it has elapsed the imposed time. The limit depth changes as a function of the diameter of the particles. The integration of all classes of diameter of the distribution of the fat globules risen up provides the estimate of the skimmed fat.
The model showed good predictive ability with R2 =0.99, even without taking into account the effects of interaction between membrane proteins of fat globules. The NIR spectra of skimmed milk and cream were used to compare the effective particle size distributions and those predicted by the rising up model. The possibility of predicting the degree of fat rising upon the basis of the characteristics of distribution of the fat globules allows to assess the actual technological impact of the dimensional variability observed in the monitoring of individual and mass milk.
G. Cabassi et al. Journal of Near Infrared Spectroscopy 21: 5 359–373