%0 Journal Article
%T 基于RBF的火龙果冷冻干燥模型预测<br>Prediction of Freeze-Drying of Pitaya Based on RBF Neural Network
%A 孟令启
%A 曾蕾
%J Hans Journal of Food and Nutrition Science
%P 162-168
%@ 2166-6121
%D 2024
%I Hans Publishing
%R 10.12677/hjfns.2024.132020
%X 以采用卧式转换型冷冻冷藏箱冻结火龙果得到的实验数据为基础，利用Matlab人工神经网络工具箱，建立了火龙果冷冻干燥的内部干燥过程与其火龙果表面压力、升华界面位移、辐射换热量、为对流换热量及变形程度对应关系的RBF神经网络预测模型。分析了变形温度和变化速度对火龙果冷冻干燥网络模型精度的影响。得出随着变化温度的增加，网络的预测误差逐渐增大；随着变形速度的增大，网络的预测误差逐渐减小的结论。通过与BP网络和Elman网络模型相比较，结果表明，RBF网络模型具有更高的精度和较强的泛化能力。<br />
Based on the experimental data obtained by freezing pitaya in a horizontal conversion type freezer refrigerator, an RBF neural network prediction model of the internal drying process of pitaya freeze-drying in relation to its surface pressure, sublimation interface displacement, radiant heat transfer, convective heat transfer and the degree of deformation was established by using the Matlab Artificial Neural Network Toolbox. The effects of deformation temperature and change rate on the accuracy of the pitaya freeze-drying network model were analyzed. It was concluded that the prediction error of the network gradually increased with the increase of change in temperature, and the prediction error of the network gradually decreased with the increase of deformation speed. By comparing with the BP network and Elman network model, the results show that the RBF network model has higher accuracy and stronger generalization ability.
%K RBF，神经网络，冷冻干燥，预测<br>RBF
%K Neural Network
%K Freeze-Drying
%K Prediction
%U http://www.hanspub.org/journal/PaperInformation.aspx?PaperID=86496