基于BP神经网络的饲料产品品质预测方法

Prediction of the quality of feed based on the BP network

饲料加工生产过程是一个非常复杂的系统,其产品品质受原料特性和加工参数的影响,且各加工参数之间相互关联,很难用精确的数学模型来反映原料特性和加工参数与产品品质的关系。利用BP神经网络能自身从样本数据中提取相关的信息,并依据数据建立模型,可以有效准确地预测产品品质。试验结果表明,产品淀粉糊化度预测结果与真值的相关系数R在0.99以上,产品水分含量预测结果与真值的相关系数R在0.97以上,预测效果理想。因此,BP神经网络预测模型用于饲料产品加工品质预测是可行的。

The production of feed is a complex process and the quality of products is affected by the characteristic of material and parameters of process while these factors are related to each other.So it is difficult to use a mathematical model reflect the relation of the quality of products,characteristic of material and parameters of process.Through the test based on the predict model,it shows that the BP（Back Propagation） neural network has the capability of self-adaptive and self-learning,and it can predict the quality of feed product accurately.The result of the prediction indicates that the correlation（R） of the starch gelatinization degree predictive value and the true value is above 0.99 and the correlation（R） of the moisture content predictive value and the true value is above 0.97,so it is available to use BP neural network on the prediction of feed quality.