论文部分内容阅读
在挤出工艺中,离开机头的物料温度通常比该段机筒设定温度要高,这样的温升往往会影响制品质量和挤出性能。针对挤出温升的非线性与复杂性,文章基于Matlab软件建立了以双螺杆转速、喂料量、物料流变特性为输入,以该工艺条件下的近机头端物料实际温度与该段机筒设定温度间的温升差值为输出的BP网络模型。结果表明:训练好的BP神经网络可以很好地预测不同的配方在不同工艺条件下的温升状况,为确定合理的挤出工艺设定温度提供了理论指导。
In the extrusion process, leaving the head of the material temperature is usually higher than the set barrel temperature, this temperature will often affect the product quality and extrusion performance. According to the non-linearity and complexity of extrusion temperature rise, the article takes the dual-screw speed, feeding volume and material rheological property as input based on Matlab software, The temperature difference between barrel set temperature is the output BP network model. The results show that the trained BP neural network can well predict the temperature rise of different formulations under different process conditions, and provide theoretical guidance for determining the proper temperature for extrusion process.