基于智能集成模型前胡切片微波真空干燥过程水分比预测

目的 提出一种基于集成学习和反向传播（back propagation，BP）神经网络的智能集成建模方法，用于建立前胡切片微波真空干燥过程水分比预测模型。方法 合理设计前胡切片微波真空干燥实验方案，探究不同微波功率、切片厚度和干燥层数对前胡切片干燥过程水分比的影响，并构建用于水分比建模的原始数据集。将原始数据集合理划分为训练集、验证集和测试集，利用训练集建立不同隐含层神经元个数的BP神经网络子模型，验证集用于防止训练过程过拟合，利用测试集建立前胡切片干燥过程水分比的智能集成模型，并进行模型应用。结果 干燥实验结果表明，提高微波功率或减少切片厚度都能够有效缩短前胡切片的总干燥时间，达到提高干燥效率、节能的目的。模型应用结果表明，智能集成模型具有较好的稳定性、泛化性和预测精度，可以快速准确地实现前胡切片微波真空干燥过程水分比预测。结论 智能集成模型为前胡切片干燥过程优化与控制提供了重要的技术支持。智能集成建模方法为中药材干燥过程建模提供了一种有效的新建模方法。

Prediction of moisture ratio for Peucedani Radix slices during microwave vacuum drying based on intelligent integrated model

Objective An intelligent integrated modeling method based on ensemble learning and back propagation (BP) neural network was proposed to establish a moisture ratio prediction model for Qianhu (Peucedani Radix) slices during microwave vacuum drying. Methods The experimental scheme of Peucedani Radix slices during microwave vacuum drying was reasonably designed to explore the effects of different microwave power, slice thickness and drying layers on the moisture ratio of Peucedani Radix slices during drying process, and the original data set for moisture ratio modeling was constructed. The original data set was reasonably divided into training set, validation set and testing set. The training set was used to establish BP neural network sub-models with different numbers of hidden layer neurons, and the validation set was used to prevent over-fitting of the training process. The testing set was used to establish an intelligent integrated model of moisture ratio of Peucedani Radix slices during microwave vacuum drying, and the model was applied. Results The results of drying experiments showed that increasing microwave power or reducing slice thickness could effectively shorten total drying time of Peucedani Radix slices, and achieve the purpose of improving drying efficiency and energy saving. The application results of the model showed that the intelligent integrated model had good stability, generalization and prediction accuracy, and could quickly and accurately predict the moisture ratio of Peucedani Radix slices during microwave vacuum drying. Conclusion The intelligent integrated model provides important technical support for drying process optimization and control of Peucedani Radix slices. The intelligent integrated modeling method provides an effective new modeling method for drying process modeling of Chinese herbal medicines.