摘要
以淀粉和乙烯-醋酸乙烯共聚物(EVA)为原料制备生物质复合材料,研究了模口温度、螺杆转速及EVA含量对挤出胀大现象的影响,并应用人工神经网络技术建立双层BP神经网络模型,将正交实验结果作为样本带入模型进行训练。经反复训练,修正误差后对复合材料的挤出胀大比进行预测。结果表明,该神经网络模型能较为准确的预测复合材料的挤出胀大比;随着模口温度的升高,或螺杆转速的下降,或EVA含量的降低,复合材料的挤出胀大比逐渐减小。
With starch and ethylene/vinyl acetate copolymer （EVA） as the raw material to preparation biomass composite materials, the influence of die temperature, screw speed and EVA content on extrusion swelling ratio have been studied, and double-layer BP network model based on artificial neural network technique has been built. The orthogonal experiment result as sample to train and forecast the extrusion swelling ratio of biologic composite materials. After repeated training and error correction to estimate the extrusion swell ratio of the composite materials. The results showed that this BP network model was effective to forecast extrusion swell ratio of composite materials accurately, with the increase of die temperature, or the falling of screw speed, or the reduce of EVA content, the extrusion swell ratio of the composite materials decreases.
出处
《广东化工》
CAS
2016年第2期19-21,共3页
Guangdong Chemical Industry
关键词
BP神经网络
模口温度
螺杆转速
挤出胀大比
BP neural network
die temperature
screw speed
extrusion swelling ratio
