Rapid detection of liquefaction force of Daqu based on near infrared spectroscopy

In order to meet the demand of Daqu liquefaction force detection, near-infrared diffuse reflectance spectroscopy technology combined with chemometric methods was used to measure it. After derivative processing of the spectral data, the characteristic wavelengths of Daqu liquefaction force were screened by competitive adaptive reweighted sampling algorithm, and a quantitative detection model was established in combination with partial least squares regression. The results showed that the prediction set determination coefficient (Rp²), root mean square error (RMSEP) and relative percentage error (RPD) of the established model were 0.862, 0.120 and 2.403, respectively. This method has good accuracy, greatly reduces the amount of data while improving the model performance, and can be used for quantitative detection of Daqu liquefaction force.