Abstract
Brewing is an ancient process which started in the middle east over 10,000 years ago. The style of beer varies across the globe but modern brewing is very much the same regardless of the style. While there are thousands of compounds in beer, current methods of analysis rely mostly on the content of only several important processing parameters such as gravity, bitterness, or alcohol. Near infrared and mid infrared spectroscopy offer opportunities to predict dozens to hundreds of compounds simultaneously at different stages of the brewing process. Importantly, this is an opportunity to move deeper into quality through measuring wort and beer composition, rather than just content. This includes measuring individual sugars and amino acids prior to fermentation, rather than total °Plato or free amino acids content. Portable devices and in-line probes, coupled with more complex algorithms can provide real time measurements, allowing brewers more control of the process, resulting in more consistent quality, reduced production costs and greater confidence for the future.
Highlights
Brewing is one of man-kind’s oldest food processes
Infrared sensors calibrated for amino acids and hordeins would be the most useful of the nitrogenous compounds that could be tracked during the brewing process as a way to observe potential yeast fermentation, and foam and haze stability, respectively
Other areas where NIR has been issued in the brewing process is to assess bitterness [57,66,74,75,76], wort and beer color [57], pH [57], biomass [65] and organic acids such as citric acids [66]
Summary
Brewing is one of man-kind’s oldest food processes. The earliest time fermented beverages were produced was many thousands of years ago. This review will discuss where infrared (both near infrared (NIR) and mid-infrared (MIR)) have been used to screen for grain quality, malt quality, hop quality and within the brewing process and provide insights to where additional in-line technology would be advantageous to the industry, Appl. Predictive modelssubstances, are built onespecially, samples that have had a reference analysis out as(wort well as having in most organic plant material (barley and hops)carried and liquids and beer). Spectral (MLR), partial least squares (PLS) ordata more advanced machine learning such as regression artificial neutral data is modelled against the reference using algorithms such as multiple linear (MLR), networks (ANN) These algorithm providelearning real-time data if builtnetworks into the partial least squares (PLS). The spectra be captured and results predicted (ANN) These programs can provide real-time data can predictions if built into the instruments off-line.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
