白竣文, 周存山, 蔡健荣, 肖红伟, 高振江, 马海乐. 南瓜片真空脉动干燥特性及含水率预测[J]. 农业工程学报, 2017, 33(17): 290-297. DOI: 10.11975/j.issn.1002-6819.2017.17.038
    引用本文: 白竣文, 周存山, 蔡健荣, 肖红伟, 高振江, 马海乐. 南瓜片真空脉动干燥特性及含水率预测[J]. 农业工程学报, 2017, 33(17): 290-297. DOI: 10.11975/j.issn.1002-6819.2017.17.038
    Bai Junwen, Zhou Cunshan, Cai Jianrong, Xiao Hongwei, Gao Zhenjiang, Ma Haile. Vacuum pulse drying characteristics and moisture content prediction of pumpkin slices[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2017, 33(17): 290-297. DOI: 10.11975/j.issn.1002-6819.2017.17.038
    Citation: Bai Junwen, Zhou Cunshan, Cai Jianrong, Xiao Hongwei, Gao Zhenjiang, Ma Haile. Vacuum pulse drying characteristics and moisture content prediction of pumpkin slices[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2017, 33(17): 290-297. DOI: 10.11975/j.issn.1002-6819.2017.17.038

    南瓜片真空脉动干燥特性及含水率预测

    Vacuum pulse drying characteristics and moisture content prediction of pumpkin slices

    • 摘要: 为探索南瓜片真空脉动干燥特性,并实现干燥过程中南瓜的含水率预测,该文研究了不同常压保持时间、真空保持时间、干燥温度和切片厚度对南瓜干燥时间和速率的影响;利用温度传感器实时采集南瓜在干燥过程中的中心温度,阐述压力脉动过程对物料传热传质的影响;建立了输入层个数为5,隐藏层个数为11,输出层为南瓜含水率,结构为"5-11-1"的BP神经网络模型,实现对南瓜含水率实时预测。结果表明:真空保持时间和常压保持时间均对南瓜干燥时间有显著影响,干燥温度60 ℃,切片厚度7 mm条件下,常压保持时间10 min和真空保持时间9 min所用干燥时间最短,约为352 min;干燥温度和切片厚度均对干燥时间有显著影响,提高干燥温度、减少切片厚度能够有效缩短干燥时间。采用Levenberg-Marquardt算法为训练函数,经过有限次训练得到的BP神经网络模型,其预测值与实测值之间的决定系数R2为0.996 8,均方根误差RMSE为0.017 3,能够很好预测南瓜在真空脉动干燥过程中的含水率。研究结果为南瓜真空脉动应用以及含水率在线预测提供理论依据。

       

      Abstract: Abstract: Pumpkin is one of the most important vegetable crops grown in the world because of its nutritional qualities. Pumpkin is a kind of seasonal vegetable, which is generally harvested from August to September. Due to its high moisture content, pumpkin cannot be stored for a long time even in low temperature environment. Drying is one of the most important processing processes for pumpkin, which can greatly prolong the shelf life. At present, the most common drying method of pumpkin is hot air drying. Hot air drying is relatively simple and easy to control, but it also has some problems such as long drying time, browning seriously and loss of nutrient. Vacuum pulse drying is a new type of drying technology, which has high drying efficiency and is very suitable for heat-sensitive materials. The aim of this study was to investigate the drying characteristics and to predict the moisture content of pumpkin slices during the vacuum pulse drying process. The effect of atmospheric pressure duration (AD), vacuum duration (VD), drying temperature and slice thickness on the drying rate and drying time of pumpkin slices was studied. The effect of pressure pulsation process on heat and mass transfer of pumpkin slices was expounded on the basis of core temperatures which were gathered by temperature sensor during drying process. The BP (back propagation) neural network model was established with the architecture of "5-11-1" which included 5 input layers of AD, VD, drying temperature, slice thickness and drying time, 11 hidden layers and single output layer of moisture content. The results demonstrated that AD and VD showed a significant impact on drying time. In a certain range, the application of longer AD time and shorter VD time not only can improve the drying efficiency, but also can significantly reduce vacuum pump running time during the drying process. The shortest drying time was about 352 min with the AD of 10 min and the VD of 9 min under drying temperature of 60 ℃ and slice thickness of 7 mm. The drying temperature and slice thickness both showed a significant impact on drying rate and drying time, and higher drying temperature and thinner thickness would result in higher drying rate. Unlike the hot air or infrared drying method, the core temperature of pumpkin presented an alternate 'high-low-high-low' phenomenon. During the VD, the core temperature decreased rapidly due to the endothermic process of moisture evaporation, and after the drying chamber pressure returned to atmospheric pressure, the core temperature was raised rapidly due to the heating effect by electrical heating plate. In the early period of drying process, the change range of core temperature between AD and VD phase was greater than that of the late drying period. The BP neural network model was trained for finite iteration calculation with Levenberg-Marquardt (LM) algorithm as the training function and tansig-purelin as the network transfer function. The determination coefficient (R2) and root mean squared error (RMSE) between the predicted and measured values were 0.9968 and 0.0173, respectively. The results will provide theoretic reference and technical supports for the application of vacuum pulse drying and the on-line prediction of moisture content in pumpkin drying process.

       

    /

    返回文章
    返回