魔芋薄层干燥试验及数学模型的建立  

Experimental Study and Mathematical Modeling on Thin Layer Drying of Konjac

作　　者：吴绍锋[1] 邱兵涛[1] 彭桂兰[1] 

机构地区：[1]西南大学工程技术学院,重庆400715

出　　处：《中国粮油学报》2016年第8期105-110,共6页Journal of the Chinese Cereals and Oils Association

基　　金：西南大学博士基金(SWUB2007021)

摘　　要：在对魔芋进行三因素三水平薄层干燥试验的基础上,分析了热风温度(50、60、70℃)、风速(0.75、1.45、1.95 m/s)及芋片厚度(5、6、7 mm)对魔芋干燥速率的影响。试验表明,魔芋干燥时间随着热风温度的升高、风速的增大以及芋片厚度的减小而减少,且热风温度对魔芋干燥速率的影响最显著;用三个经典数学模型(Henderson and Pabis,Lewis和Page模型)及三次多项式模型对试验数据进行拟合,经拟合得出最适合魔芋干燥的模型为三次多项式模型;用BP神经网络建立魔芋干燥的数学模型,并与三次多项式模型的拟合结果进行对比,结果表明,BP神经网络拟合的含水率比平均相对误差为0.94%,精度明显高于三次多项式模型(5.64%)。This paper discussed the effects of air temperature （50,60 ℃ and 70℃ ） , air velocities （0.75,1.45 m/s and 1.95 m/s）and the thickness of konjac pieces（5,6 mm and 7 ram）on the drying rate of konjac. The thin layer drying experiments of konjac were conducted through three - factor and three - level orthogonal experimental de- sign. Tests showed the drying time decreased with the increase of the hot air temperature and air velocities and the decrease of the thickness of konjac slices. The experiments also showed that the air temperature was the main influen- tial factor in the drying process. Three widely recommended mathematical models （Henderson and Pabis, Lewis and Page models）and the Cubic model were selected to fit the experimental data, and the Cubic model was much more ad- equate model for describing the thin layer drying of konjac. The mathematical model of konjac drying based on BP neural network was also established. Through fitting the experimental data using the BP neural network model and the Cubic model respectively, the results indicated that the moisture ratio predicted by the BP neural network model（ aver- age relative error was 0.94% ）was more accurate than that predicted by the Cubic model（5.64% ）.

关 键 词：魔芋 薄层干燥 数学模型 BP神经网络 

分 类 号：S375[农业科学—农产品加工]