科研速览 · Science Skim继续刷下去 · Keep skimming →
◆ Microchemical Journal2026-01-17· Partial least squares regression

Multi-component beer quality control using miniaturized Fourier transform near-infrared systems

Haona Bao, Luis Rodriguez‐Saona

原始摘要(英文原文)· Original abstract
Traditional methods for assessing beer quality are often time-consuming and require multiple instruments and trained personnel, posing significant challenges particularly for microbreweries. This study investigates the feasibility of miniaturized Fourier transform near-infrared (FT-NIR) spectroscopy combined with multivariate data analysis as a cost-effective solution for simultaneously assessing multiple quality attributes. Spectral data were acquired using two portable devices, a self-assembled FT-NIR (Hamamatsu C15511-01, 1100–2500 nm) and a commercial FT-NIR (NeoSpectra Scanner, 1350–2550 nm). A total of 132 beer samples with a wide range of styles were collected from local breweries and liquor stores. Six key quality parameters, including color, specific gravity, real extract, alcohol level (ABV%), pH and total isoalpha acids content (IAA) were determined using reference methods from the American Society of Brewing Chemists to develop spectroscopic analysis models based on Partial Least Squares Regression (PLSR). Both sensors show excellent prediction capability for ABV% with coefficient correlation (R PRE ) of 0.99 and standard error of prediction (SEP) lower than 0.3%. The self-assembled FT-NIR device with wider spectral range and higher resolution showed better prediction performance in real extract (R PRE = 0.97, SEP = 0.27°Plato), specific gravity (R PRE = 0.95, SEP = 0.0011), color (R PRE = 0.96, SEP = 1.15 SRM), pH (R PRE = 0.92, SEP = 0.059), and total IAA content (R PRE = 0.88, SEP = 1.76 ppm) and ratio of Performance to Deviation (RPD) above 2.5. These results confirm the potential of miniaturized FT-NIR spectroscopy as a reliable alternative to conventional methods for rapid beer quality assessments. • Two handheld FT-NIR devices were evaluated for beer analysis. • Six key quality attributes of beer were assessed across diverse styles. • Robust and accurate PLSR predictive models were built from NIR spectra. • Miniaturized FT-NIR enabled simple, rapid, multi-parameter assessments.
读原文 · Read the paper ↗

AI 追问PRO

登录后使用 AI 追问

讨论区

登录后参与讨论

相关论文 · Related

Multi-component beer quality control using miniaturized Fourier transform near-infrared systems — 科研速览 Science Skim