Quantitative analysis of key components in Qingke beer brewing process by multispectral analysis combined with chemometrics.

By sampling the wort from different stages of mashing and boiling, part of the sample was used for the acquisition of three spectra (NIR, Raman and UV–Vis) and the other part for chemical analysis. The results of chemical analysis and spectra were combined using chemometrics, and three models of PLS((Partial Least Squares), LS-SVM(Least Squares Support Vector Machines), and NN(Neural Networks) were developed to quantify reducing sugar, free amino nitrogen, and total phenols contents in wort. The aim of the study is to investigate the best spectra and models for the quantitative analysis of key wort components during mashing and boiling of Qingke malt, so as to achieve real-time monitoring of the Qingke beer brewing process. [Display omitted] • Qingke malt has the potential to become a raw material for beer brewing. • Quantitative models were developed using three spectra in combination with chemometrics. • The study focuses on the brewing process of beer to make it clearer and more controllable. In order to monitor the Qingke beer brewing process in real time, this paper presents an analytical method for predicting the content of key components in the wort during the mashing and boiling stages using multi-spectroscopy combined with chemometrics. The results showed that the Neural Networks (NN) model based on Raman spectroscopy (RPD = 3.9727) and the NN model based on NIR spectroscopy (RPD = 5.1952) had the best prediction performance for the reducing sugar content in the mashing and boiling stages; The partial least Squares (PLS) model based on Raman spectroscopy (RPD = 2.7301) and the NN model based on Raman spectroscopy (RPD = 4.3892) predicted the content of free amino nitrogen best; The PLS model based on UV–Vis spectroscopy (RPD = 4.0412) and the NN model based on Raman spectroscopy (RPD = 4.0540) are most suitable for the quantitative analysis of total phenols. The results can be used as a guide for real-time control of wort quality in industrial production. [ABSTRACT FROM AUTHOR]