基于BP神经网络的旁热式辐射与对流粮食干燥过程模型

发布时间：2018-03-22 14:15

本文选题：粮食　切入点：红外辐射与对流干燥　出处：《农业机械学报》2017年03期 　论文类型：期刊论文

【摘要】：针对旁热式辐射与对流粮食干燥机的干燥特点,建立了一种粮食干燥机干燥过程的BP神经网络预测模型。该模型采用了3层神经网络结构(8-10-1),模型输入为粮食干燥机的8个变量,模型输出为出口粮食水分比或干燥速率。通过编写Matlab建模程序,基于实际干燥实验的样本数据训练与测试网络,实现了红外辐射与对流联合干燥的动力学模型,并给出了相应的模型数学表达式,模型预测的出口水分比与干燥速率的R2分别为0.998 9和0.998 0,均方根误差分别为0.009和0.004 1,预测结果与实际测量数据拟合较好;另外,结合实验干燥条件对模型干燥性能的预测结果进行了分析与总结,并依据同样方法建立了顺逆流粮食干燥过程的出口粮食水分比预测模型,对比了2种干燥方式的干燥性能。仿真预测表明用BP神经网络方法建模简单,具有自适应性、灵活性和自学习性等特点,相比于其他粮食干燥的经验数学模型,能综合考虑多种影响因素,可为红外辐射与对流联合干燥过程提供一种新的建模方法。
[Abstract]:According to the drying characteristics of the side heat radiation and convection grain dryer, a BP neural network prediction model for the drying process of the grain dryer is established. The model adopts a 3-layer neural network structure of 8-10 ~ (-1), and the model is inputted as 8 variables of the grain dryer. The output of the model is the ratio of water to water or the drying rate of grain for export. The dynamic model of infrared radiation and convection combined drying is realized by writing Matlab modeling program and training and testing network based on the sample data of actual drying experiment. The mathematical expressions of the model are given. The R2 of the outlet water ratio and the drying rate predicted by the model are 0.998 9 and 0.998 0, respectively, and the root mean square error are 0.009 and 0.004 1, respectively. The predicted results fit well with the actual measured data. Combined with the experimental drying conditions, the prediction results of the drying performance of the model were analyzed and summarized. According to the same method, a prediction model of grain moisture ratio for export grain was established in the process of countercurrent grain drying. The simulation results show that the BP neural network method is simple in modeling, self-adaptive, flexible and self-learning, and compared with other empirical mathematical models of grain drying. A new modeling method can be provided for the combined drying process of infrared radiation and convection.
【作者单位】：北京邮电大学自动化学院;青岛农业大学理学与信息学院;中国农业大学工学院;北京联合大学自动化学院;
【基金】：国家粮食公益性行业科研专项(201413006)
【分类号】：TP183;S226.6

【参考文献】