SIA OpenIR  > 数字工厂研究室
基于迁移学习和知识蒸馏的加热炉温度预测方法
Alternative TitlePrediction method of furnace temperature based on transfer learning and knowledge distillation
翟乃举1,2,3,4; 周晓锋1,2,3; 李帅1,2,3,4; 史海波1,2,3
Department数字工厂研究室
Source Publication计算机集成制造系统
ISSN1006-5911
2021
Pages1-13
Contribution Rank1
Funding Organization辽宁省“兴辽英才计划”资助项目(XLYC1808009)
Keyword加热炉 迁移学习 时间卷积网络 知识蒸馏
Abstract

实现对加热炉炉温的精准预测以便于采用精确的控制策略对加热炉的燃烧进行优化控制,是冶金企业中燃烧装置优化控制的核心问题。现阶段少有研究关注加热炉内所有加热区的温度预测以及神经网络在炉温预测方面的适用性难题。针对此问题,提出基于迁移学习和知识蒸馏的炉温预测方法。该方法首先建立基于时间卷积网络的源域温度预测模型,然后采用生成对抗损失进行域自适应来完成模型迁移,实现所有加热区温度的准确预测。进一步建立基于多任务学习的蒸馏网络,该网络通过教师辅助学生的方式来解决深度迁移网络延时高的缺点。实验结果表明,提出的迁移学习网络可以明显提升炉温预测的准确性,蒸馏网络可以明显减少网络参数,极大的提高炉温预测时效性。

Other Abstract

It is the core problem of combustion device optimization control in metallurgical enterprises to realize the accurate prediction of furnace temperature so as to use accurate control strategy to optimize the combustion control of heating furnace. At present, few researches focus on the temperature prediction of all heating zones and the applicability of neural network in furnace temperature prediction. To solve this problem, a prediction method of furnace temperature based on transfer learning and knowledge distillation is proposed. Firstly, the temperature prediction model of the source zone based on the temporal convolution network is established, and then the model transfer is completed by using the generative adversarial loss for domain adaptation, so as to realize the accurate prediction of the temperature in all heating zones. A distillation network based on multi-task learning is further established, which solves the shortcomings of the high delay of the deep transfer network by the way of teacher assisting student. The experimental results show that the proposed transfer learning network can significantly improve the accuracy of furnace temperature prediction, distillation network can significantly reduce network parameters, greatly improve the timeliness of furnace temperature prediction.

Language中文
Document Type期刊论文
Identifierhttp://ir.sia.cn/handle/173321/28118
Collection数字工厂研究室
Corresponding Author周晓锋
Affiliation1.中国科学院网络化控制系统重点实验室
2.中国科学院沈阳自动化研究所
3.中国科学院机器人与智能制造创新研究院
4.中国科学院大学
Recommended Citation
GB/T 7714
翟乃举,周晓锋,李帅,等. 基于迁移学习和知识蒸馏的加热炉温度预测方法[J]. 计算机集成制造系统,2021:1-13.
APA 翟乃举,周晓锋,李帅,&史海波.(2021).基于迁移学习和知识蒸馏的加热炉温度预测方法.计算机集成制造系统,1-13.
MLA 翟乃举,et al."基于迁移学习和知识蒸馏的加热炉温度预测方法".计算机集成制造系统 (2021):1-13.
Files in This Item:
File Name/Size DocType Version Access License
基于迁移学习和知识蒸馏的加热炉温度预测方(1042KB)期刊论文出版稿开放获取CC BY-NC-SAView Application Full Text
Related Services
Recommend this item
Bookmark
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[翟乃举]'s Articles
[周晓锋]'s Articles
[李帅]'s Articles
Baidu academic
Similar articles in Baidu academic
[翟乃举]'s Articles
[周晓锋]'s Articles
[李帅]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[翟乃举]'s Articles
[周晓锋]'s Articles
[李帅]'s Articles
Terms of Use
No data!
Social Bookmark/Share
File name: 基于迁移学习和知识蒸馏的加热炉温度预测方法.pdf
Format: Adobe PDF
All comments (0)
No comment.
 

Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.