RESUMO
The objective of this work was to devise a complete procedure based on chemometrics and the use NIR spectroscopy at the entrance of a feed mill to provide early evidence of non-conformity and unusual ingredients and thus help to achieve cost-savings. The procedure was validated at laboratory level and was adapted for application at the Cargill Animal Nutrition feed mill. The study focused on the characterisation of pure soybean meal with the aim of creating an early control system for detecting and quantifying any unusual ingredient that might be present in the soybean meal, such as melamine, cyanuric acid or whey powder (milk serum). The study results showed that the use of NIR, combined with some simple chemometric tools based on distances and residuals from regression equations, is appropriate for authenticating important feed products (in this case, soybean meal) and detecting the presence of abnormal samples or impurities in both the laboratory and at the feed mill.