Scholarlink

<p>This study built a prediction model to achieve the rheological properties of Chinese dates and donkey-hide gelatin mixed paste. Spectral data were collected from the samples with near-infrared hyperspectral imaging (NIR-HIS) to detect their moisture online. The moisture content prediction model constructed on the partial least squares (PLS) regression of standard normal variate (SNV) preprocessing and synergy interval partial least squares (siPLS) characteristic wavelength selection shows the optimal performance (R²=0.9783, RPD = 6.7813, RMSEP = 0.0192). Next, rheological tests were conducted on samples with different moisture contents. The prediction model on the flow index (n) and consistency parameter (K) established by the voting regression algorithm shows the optimal performance. (n: R²<i>p</i> = 0.9000, RPD = 3.1628, RMSEP = 0.0226; K: R²<i>p</i> = 0.9246, RPD = 33.64, RMSEP = 440.1094). In summary, NIR-HIS successfully predicted the flow index and consistency parameter of the food mixture. This enabled the calculation of shear stress, apparent viscosity, and stress-strain relationships. This offers a non-destructive, fast, and direct assessment of rheological properties that cannot be replicated by conventional rheometers. It allows for real-time monitoring of the mixing state, helping to avoid quality issues caused by inadequate mixing or excessive energy input from over-mixing.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Prediction model for rheological properties of Chinese dates and donkey-hide gelatin mixed paste based on near-infrared hyperspectral imaging

  • 期刊论文

  • 发表日期：

    2025年6月27日

查看全文
  • Xiaotong Pang,
  • Yangyang Su,
  • Qilong Xue,
  • Zhouyou Wu,
  • Weiqiang Huang,
  • Zhongchun Ma,
  • Shouwang Wang,
  • Yang Yu,
  • Zheng Li

摘要

This study built a prediction model to achieve the rheological properties of Chinese dates and donkey-hide gelatin mixed paste. Spectral data were collected from the samples with near-infrared hyperspectral imaging (NIR-HIS) to detect their moisture online. The moisture content prediction model constructed on the partial least squares (PLS) regression of standard normal variate (SNV) preprocessing and synergy interval partial least squares (siPLS) characteristic wavelength selection shows the optimal performance (R²=0.9783, RPD = 6.7813, RMSEP = 0.0192). Next, rheological tests were conducted on samples with different moisture contents. The prediction model on the flow index (n) and consistency parameter (K) established by the voting regression algorithm shows the optimal performance. (n: R²p = 0.9000, RPD = 3.1628, RMSEP = 0.0226; K: R²p = 0.9246, RPD = 33.64, RMSEP = 440.1094). In summary, NIR-HIS successfully predicted the flow index and consistency parameter of the food mixture. This enabled the calculation of shear stress, apparent viscosity, and stress-strain relationships. This offers a non-destructive, fast, and direct assessment of rheological properties that cannot be replicated by conventional rheometers. It allows for real-time monitoring of the mixing state, helping to avoid quality issues caused by inadequate mixing or excessive energy input from over-mixing.