Near Infrared Spectroscopy and Multivariate Analysis methods for monitoring flour performance in an industrial bread-making process

The present study is aimed at evaluating the possibility to predict bread speci\ufb01cations, for an industrial bread-making process, on the basis of the properties of \ufb02our employed in production. The \ufb02our delivered at the production plant, of which rheological and chemical properties were available, were analysed by means of Near Infrared Spectroscopy (NIRS). Based on the \ufb02our properties and NIR signals, multivariate control charts were constructed in order to detect \ufb02our batches leading to a bread with non-optimal behaviour. The results show that it is possible to distinguish \ufb02our batches leading to a product with a particularly negative performance, by modelling the properties commonly measured on \ufb02ours and the acquired Near Infrared signals. In spite of the absence of monitoring of process variables, which could have offered a more sound basis for the interpretation, especially when false positives and negatives are detected, these results are of particular interest from the point of view of raw material evaluation in process monitoring. Also, the potentiality of Near Infrared Spectroscopy allows considering this approach for an on-line implementation in the control of incoming raw materials in this industrial process