基于人工神经网络的落叶松干燥模型研究  

Study on the Drying Model of Larch Wood Based on Artificial BP Neural Network

作　　者：谢键[1] 王晓丰[1] 段文英[1] 陈广元[1] 

机构地区：[1]东北林业大学理学院,哈尔滨150040

出　　处：《森林工程》2013年第2期58-61,共4页Forest Engineering

摘　　要：采用人工神经网络BP型3层映射模式,对东北林业大学木材干燥实验室俄罗斯产落叶松进行木材含水率测定,干球温度平衡含水率和预热阶段干燥阶段的木材含水率以干燥机内部所测定为基准作为输入矩阵,以所测定的木材降温阶段和湿热阶段所测定的木材含水率作为输出矩阵从而确定3层形式,作出一段周期内落叶松控制系统降温和湿热阶段含水率预测,通过网络训练获得最佳权值,作为预测模拟参数,通过调整干球温度和其他参数使在半自动控制中木材含水率达到可控效果,从而在今后的干燥过程中可以通过含水率的预测得以实现对温度的调控。This paper aims to determine the moisture content of Russian larch wood in the wood drying lab of Northeast Forestry University by using the three-level mapping mode of artifieial BP neural network. The equilibrium moisture contents at dry-bulb temperature, pre-heating period, and drying phase, which are obtained from the interior of the drying machine, were used as input matrix. The measured moisture contents during cooling phase and hot-humid phase were the output matrix. Thus, a lareh wood moisture prediction model at cooling and hot-humid stages was established. Through network training, the best weights of the parameters in the forecasting model were obtained. By modifying dry-bulb temperature and other parameters, the controllable effect of semi-automation wood moisture content can be achieved, so that people may control the temperature by predicting the moistm＇e content of wood in the future drying process.

关 键 词：BP神经网络 落叶松木材 干燥 木材含水率 

分 类 号：S757.3[农业科学—森林经理学]