为了快速监测干燥过程中水分和油脂的变化，采用低场核磁共振（LF-NMR）技术对恒温和变温滚筒催化红外-热风干燥过程中核桃自由水、弱和强结合水及油脂进行测定，并与单一热风干燥相比较，同时建立水分横向弛豫峰面积占比与含水率之间的数学模型。结果表明，三种干燥方式下的干燥曲线变化趋势一致，与单一热风干燥相比[时长20 h，最大干燥速率0.11 g/(g·min)]，变温和恒温红外-热风干燥效率[时长14.38 h和16.16 h，最大干燥速率0.28 g/(g·min)和0.48 g/(g·min)]显著提高，干燥时间分别缩短了28.10%和19.20%，最大干燥速率分别提升了2.55和4.36倍，表明红外干燥处理显著提升了干燥效率。LF-NMR横向弛豫图谱显示干燥过程中自由水峰面积显著下降直至消失，弱结合水峰面积显著降低，强结合水和油脂峰面积基本不变，表明干燥处理去除了几乎全部的自由水和大部分的弱结合水，对强结合水和油脂含量无明显影响。建立了水分弛豫峰面积占比与实际含水率之间的数学模型，R2均大于0.90，预测效果良好。总之，与国标法相比，LF-NMR能够很好地对干燥过程中水分和油脂变化进行检测，是一种高效快速检测含水率的新型方法。

In order to quickly monitor the changes of moisture and oil in the drying process, low field-nuclear magnetic resonance (LF-NMR) technology was used to determine the free water, weakly and strongly bound water and oil of walnuts under the constant temperature and variable temperature drum catalytic infrared-hot air drying, and compared with the single hot air drying. Moreover, a mathematical model between the proportion of the moisture lateral relaxation peak area and the actual moisture content was established. The results showed that the changing trend of the drying curve under the three drying methods was the same. Compared with the single hot air drying [duration 20 h, maximum drying rate 0.11 g/(g·min)], the variable temperature and constant temperature drum catalytic infrared-hot air drying efficiencies [drying times of 14.38 h and 16.16 h, maximum drying rates of 0.28 g/(g·min) and 0.48 g/(g·min)] were significantly increased, the drying times were shortened by 28.10% and 19.20% respectively, and the maximum drying rates were increased by 2.55 and 4.36 times respectively, indicating that the catalytic infrared drying treatment significantly improved drying effectiveness. The LF-NMR transverse relaxation map showed that the free water peak area decreased significantly during the drying process until it disappeared, the weakly bound water peak area was significantly reduced, and the strongly bound water and oil peak areas remained basically unchanged, indicating that the drying process removed almost all the free water and most of the weakly bound water and had no obvious effects on the strongly bound water and oil contents. A mathematical model between the proportion of the moisture relaxation peak area and the actual moisture content was established. R2 was greater than 0.90, and the prediction effect was good. In general, compared with the national standard method, LF-NMR can detect the change of moisture and oil during the drying process very well, and it was a new method for efficient and rapid detection of moisture content.

国家重点研发计划政府间国际科技创新合作重点专项（2017YFE0105300）