On-line prediction of cheese making indices using backscatter of near infrared light

The potential of a fibre optic sensor, detecting light backscatter in a cheese vat during coagulation and syneresis, to predict curd moisture, fat loses and curd yield was examined. Temperature, cutting time and calcium levels were varied to assess the strength of the predictions over a range of processing conditions. Equations were developed using a combination of independent variables, milk compositional and light backscatter parameters. Fat losses, curd yield and curd moisture content were predicted with a standard error of prediction (SEP) of ±2.65 g 100 g−1 (R2=0.93), ±0.95% (R2=0.90) and ±1.43% (R2=0.94), respectively. These results were used to develop a model for predicting curd moisture as a function of time during syneresis (SEP=±1.72%; R2=0.95). By monitoring coagulation and syneresis, this sensor technology could be employed to control curd moisture content, thereby improving process control during cheese manufacture.