中文摘要：

从全国21个省份收集78个代表性酒糟样品,采用国标方法测定了其成分含量并进行了统计分析,结果显示粗纤维、粗灰分和粗蛋白3种成分质量分数分布范围分别为1.60%～37.54%、0.74%～26.85%和12.01%～38.69%,标准偏差分别为8.33%、5.21%和7.11%,样品成分含量差异较大。利用NIRS建立了其粗纤维、粗灰分和粗蛋白质量分数的定量分析模型,定标决定系数分别为0.98、0.91和0.96,定标标准误差分别为1.19%、1.58%和1.61%,验证决定系数分别为0.98、0.92和0.96,预测标准误差分别为1.20%、1.57%和1.60%,相对分析误差分别为7.38、3.75和4.98,模型预测精度较高,可以用于实际检测分析。

英文摘要：

Dried distillers＇ grains were rich in nutrition, which used as protein feed materials, and could give some benefit to supplement feed deficit and reduce feed cost. Totally 78 dried distillers＇ grains samples were collected from 21 provinces of China. Crude fiber, crude ash and crude protein contents were detected in laboratory by standard methods. The statistic analysis showed that the distribution range of crude fiber, crude ash and crude protein contents were 1.60% - 37.54%, 0.74% -26. 85% and 12.01% - 38.69% , respectively, the standard distribution value of them were 8.33% , 5.21% and 7.11% , respectively. The big differences in components content were mainly caused by the remaining rice hull during drying and separating process. Quantitative evaluation models were established to determinate crude fiber, crude ash and crude protein content of dried distillers＇ grains by NIRS methods. The results of NIRS analysis showed that this method could provide rapidly quantitative prediction for crude fiber, crude ash and crude protein contents in dried distillers＇ grains, the coefficient of determination in calibration were 0.98, 0.91 and 0.96, respectively; the standard error in calibration were 1. 19% , 1.58% and 1.61%., respectively; the coefficient of determination in validation were 0.98, 0.92 and 0.96, respectively; the standard error in validation were 1.20% , 1.57% and 1.60% , respectively, and the ratio of performance to standard deviation were 7.38, 3.75 and 4.98, respectively. The prediction values of these models are accurate.