基于BP神经网络预测红外-喷动干燥带壳鲜花生水分比

为实现带壳鲜花生红外-喷动干燥过程中水分比的预测，本实验探究了不同干燥温度（55、60、65 ℃和70 ℃）、进口风速（16、17、18 m/s和19 m/s）和助流剂质量（1.0、1.5、2.0 kg和2.5 kg）对带壳鲜花生干燥时间和干燥速率的影响，建立了输入层为干燥温度、进口风速、助流剂质量和干燥时间，隐含层节点数为11，输出层为带壳鲜花生水分比，拓扑结构为“4-11-1”的BP神经网络模型。结果表明：干燥温度和进口风速是影响带壳鲜花生水分比的主要因素，增加进口风速和提高干燥温度能有效缩短带壳鲜花生的干燥时间，提高干燥效率。采用Levenberg-Marquardt（L-M）算法为训练函数，选择tansig-purelin为网络传递函数，经过有限次训练得到的BP神经网络模型，其水分比预测值与实验值之间的决定系数R2为0.99，均方误差为0.02，水分比预测结果相较于传统经典数学模型准确且迅速。本研究建立的BP神经网络模型可为带壳鲜花生在红外-喷动干燥过程中的水分比在线预测提供理论依据和技术支持。

Backward Propagation (BP) Neural Network-Based Prediction of Moisture Ratio of Fresh In-shell Peanut during Infrared-Assisted Spouted Bed Drying

In order to predict the moisture ratio of in-shell peanut during infrared-assisted spouted bed drying, the effects of drying temperature (55, 60, 65 and 70 ℃), inlet airflow rate (16, 17, 18 and 19 m/s) and the amount of glidant (1.0, 1.5, 2.0 and 2.5 kg) on the drying time and drying rate were investigated. A BP neural network model whose topological structure was “4-11-1” was established using drying temperature, inlet airflow rate, the amount of glidant and drying time as the input layers, and the moisture ratio of peanut as the output layer. The number of nodes in the hidden layer of the established network was 11. The results showed that drying temperature and inlet airflow rate were the main factors affecting the moisture ratio of fresh in-shelled peanut. Increasing the inlet airflow rate or drying temperature could effectively shorten the drying time and improve drying efficiency. Using the Levenberg-Marquardt (LM) algorithm and tansig-purelin as the training function and the network transfer function, respectively, the BP neural network model was obtained after finite training. The determination coefficient (R2) between the predicted value and the experimental value was 0.99, and the mean square error was 0.02. Compared with the traditional classical model, the BP neural network model allowed rapid and accurate prediction of the moisture ratio. This model can provide a theoretical basis and technical support for on-line prediction of moisture ratio of fresh in-shell peanut during infrared-assisted spouted bed drying.