油茶籽热泵干燥特性及模型的研究  

Drying Characteristics and Models for Heat Pump Drying of Camellia Oleifera Seed

作　　者：吴雪辉[1,2] 龙婷 王泽富[1] 蓝梧涛[1] 

机构地区：[1]华南农业大学食品学院,广州510642 [2]广东省油茶工程技术研究中心,广州510642

出　　处：《中国粮油学报》2018年第1期111-117,共7页Journal of the Chinese Cereals and Oils Association

基　　金：林业公益性行业科研专项(201504703);广东省科技计划(2016A020209006)

摘　　要：为设计、制造油茶籽热泵干燥机提供理论依据,研究不同温度和相对湿度条件下油茶籽的热泵干燥特性,采用7个常用的干燥数学模型和BP神经网络模型对油茶籽热泵干燥过程进行拟合。结果表明,油茶籽的热泵干燥分加速和降速两个阶段,以降速阶段为主,温度越高,相对湿度越低,干燥时间越短,有效水分扩散系数越大,相对湿度为10%和50%时,50、60、70℃下干燥至干基含水量2%所需的时间、最大干燥速率、有效水分扩散系数分别为:21.5、22.5 h,14、17 h,10、14 h,14.0、13.8 g·(g·h)^(-1),18.6、17.0 g·(g·h)^(-1),25.0、20.1 g·(g·h)^(-1),2.830×10^(-9)、2.809×10^(-9)m^2/s,3.784×10^(-9)、3.470×10^(-9)m^2/s,6.503×10^(-9)、4.963×10^(-9)m^2/s;油茶籽热泵干燥所需的活化能为32.0 kJ/mol,从油茶籽中去除1 kg水分需要的最低能量是1 778 kJ,耗电约0.49 kW·h。常用的7种干燥数学模型中Midilli模型的平均R^2值最大、平均RMSE值、SSE值和平均χ~2最小,拟合度最高,所得理论值和实验值之间的平均相对误差为8.97%;而BP神经网络模型在训练、校验、测试过程中,模型拟合度均达0.999以上,仿真结果与实验值间的平均相对误差为4.57%,说明BP神经网络模型能更好的表征油茶籽热泵干燥过程,更准确的预测油茶籽热泵干燥过程中含水量的变化规律。In order to provide theoretical basis for the design and manufacture of heat pump dryer for the camellia oleifera seed, the characteristics of heat pump drying for camellia oleifera seeds under different temperature and relative humidity were studied. Drying process was fitted by seven kinds of drying mathematical models and the BP neural network model. The results showed that the process could be divided into 2 stages, i. e. the first stage was acceleration and the second stage was deceleration which was the main stage. The higher the drying temperature and the lower relative humidity, the shorter drying time and the greater effective water diffusion coefficient would be. When the relative humidity was 10% ,with the temperature rising from 50℃ to 70 ℃ ,the maximum drying rate was 14.0,18.6 and 25.0 g · （ g · h）^ -1 ;the drying time was 21.5,14 and 10 h;the effective water diffusion coefficient was 2. 830 × 10^-9 m^2/s,3. 784 × 10^-9 m^2/s,6. 503 × 10^-9 m^2/s, respectively. When the relative humidity was 50%, with the temperature rising from 50 -C to 70 -C ,the maximum drying rate was 13.8,17.0,20.1 g · （g · h） ^-1 ;the drying time was 22.5,17,14 h ;the effective water diffusion coefficient was 2. 809 × 10 ^-9 m^2/s ,3. 470 × 10^ -9 m^2/s ,4. 963 × 10^-9 m^2/s,respectively. The activation energy of camellia oleifera seed was 32.01 kJ/mol,which indicated the mini- mum energy required to remove 1 kg of water from camellia oleifera seed was 1 778 kJ and the consumption of power was about 0.49 kW · h. The average R^2 value of the Midilli model was the highest, the average RMSE value, SSE value and χ^2 value were the smallest, and the fitness was the highest. The average relative error between the theoretical value and the experimental value was 8.97 %. The fitting degree of BP neural network model was more than 0. 999 in the process of training, check and test, and the average relative error between the simulation result and the experimental value was 4.57% ,which indicated the BP neural netw

关 键 词：油茶籽 干燥温度 相对湿度 热泵 数学模型 BP神经网络 

分 类 号：TS222.1[轻工技术与工程—粮食、油脂及植物蛋白工程] TS223.2[轻工技术与工程—食品科学与工程]