LAUSR

Abstract In this proof-of-concept study, we consider the problem of finding the conditions at which an industrial wheat milling process should be operated to provide a product with assigned quality from a given wheat variety. We propose an approach where mathematical models are used to assist the miller in determining the required operating conditions. The basic idea we explore is to move most of the experimentation from the industrial-scale equipment to a small-scale one where data can be obtained more quickly and at an inferior cost. The operation of an industrial roller mill, whose milling gap is the operating variable to be determined, is used as a test bed to assess the feasibility of the proposed methodology. An extended set of data deriving from experimentation in a small-scale mill are used jointly with a dataset from the industrial mill to design a multivariate statistical model based on latent variables, which is then inverted to find the desired operating conditions. The results show that the proposed approach is viable and is particularly effective when the industrial-scale dataset is limited. It is also shown that near infrared spectra can be effectively used to characterize the wheat feed.