基于自编码器的切粒机熔体异常辨识研究
Research on Melt Anomaly Identification of Pelletizer Based on Auto-encoder
张灵修 1乔成斌 1朱明祥2
作者信息
- 1. 中国石化仪征化纤有限责任公司,江苏 扬州 211900
- 2. 南京师范大学泰州学院 电力工程学院,江苏 泰州 225300
- 折叠
摘要
文章对聚酯切粒机熔体挤出过程中熔体异常的形态特征和产生原因进行了分析,讨论了基于自编码器的深度学习方法在熔体异常辨识领域应用的可行性,并利用Deep Learning Tool平台研究了基于自编码器的图像异常辨识网络的应用效果.实验结果表明,文章所采用的网络具有良好的熔体异常辨识准确率和泛化性能,并初步实现了对熔体异常程度的有效区分.
Abstract
The morphological characteristics and causes of melt abnormality in the melt extrusion process of polyester pelletizer are analyzed.The feasibility of the application of deep learning method based on auto-encoder in the field of melt anomaly identification is discussed.The application effect of image anomaly identification network based on auto-encoder is studied by using Deep Learning Tool platform.The experimental results show that the network used in this paper has good melt anomaly identification accuracy and generalization performance,and initially realizes the effective distinction of melt anomaly degree.
关键词
切粒机/熔体异常辨识/自编码器/深度学习Key words
pelletizer/melt anomaly identification/autoencoder/deep learning引用本文复制引用
基金项目
泰州市科技支撑计划(社会发展)项目(SSF20230014)
南京师范大学泰州学院科研团队建设项目()
出版年
2024
NETL
NSTL
万方数据