中国塑料 ›› 2019, Vol. 33 ›› Issue (8): 69-75.DOI: 10.19491/j.issn.1001-9278.2019.08.012

• 加工与应用 • 上一篇    下一篇

基于BP神经网络的颜料分散性能预测研究

胡娜,李翱,付云松,王若寒   

  1. 北京化工大学
  • 收稿日期:2019-04-08 修回日期:2019-04-17 出版日期:2019-08-26 发布日期:2019-08-26

Study on Dispersion Performance Prediction of Pigment Based on BP Neural Network

  • Received:2019-04-08 Revised:2019-04-17 Online:2019-08-26 Published:2019-08-26

摘要: 基于正交试验,对塑料制品色差的影响因素进行分析探讨,并应用反向传播(BP)神经网络的数据预测功能,构建颜料分散性能的预测模型。结果表明,塑料制品色差影响因素主次顺序为:转子转速>混合时间>混合温度;提高转子转速,有利于颜料粒子在塑料基体中的分散和分布过程,制品着色品质得以提升;构建的BP神经网络模型预测相对误差不超过10 %,能够较好地预测塑料着色工艺中颜料的分散性能。

Abstract: Based on the orthogonal experiment, the influence factors of color difference of plastic products were analyzed and discussed, and the data-prediction function of BP neural network was utilized to construct the prediction model of pigment dispersion performance. The results indicated that the factors affecting the color difference of plastic products were listed in an order of rotor speed > mixing time > mixing temperature. The increase of rotor speed could facilitate the dispersion and distribution of pigment particles in the plastic matrix, and the coloring quality of the product was improved accordingly. The relative error predicted by the BP neural network model did not exceed 10 %, and therefore this model could predict the dispersion performance of pigments in the plastic coloring process.