基于DL-BiGRU多特征融合的注塑件尺寸预测方法

doi:10.3785/j.issn.1008-973X.2024.03.021

浙江大学学报(工学版)

2024, Vol. 58

Issue (3): 646-654 DOI: 10.3785/j.issn.1008-973X.2024.03.021

机械工程

基于DL-BiGRU多特征融合的注塑件尺寸预测方法

钱庆杰1(

),余军合1,*(

),战洪飞1,王瑞1,胡健2

1. 宁波大学机械工程与力学学院，浙江宁波 315211
2. 中机中联工程有限公司第一工业设计研究院，重庆 400039

Dimension prediction method of injection molded parts based on multi-feature fusion of DL-BiGRU

Qingjie QIAN1(

),Junhe YU1,*(

),Hongfei ZHAN1,Rui WANG1,Jian HU2

1. Faculty of Mechanical Engineering and Mechanics, Ningbo University, Ningbo 315211, China
2. The First Industrial Design and Research Institute of CMCU Engineering Co. Ltd, Chongqing 400039, China

全文: PDF(1372 KB) HTML

摘要：

为了充分挖掘注塑成型过程中模腔内的高频时序特征和注塑成型机状态特征，提出基于双层双向门控循环单元网络(DL-BiGRU)的多特征融合注塑件尺寸预测方法. 分析膜腔内传感器高频时序特征与注塑件尺寸间的关联性，采用DL-BiGRU网络从高频数据中自动提取时序特征，表征注塑件成型过程状态变化特性. 通过采样模腔内高频时序数据进行展成平铺，表征注塑成型的瞬时特征. 融合时序特征、瞬时特征和成型机状态特征，构建端到端的深度学习多特征融合框架. 将上述3种特征融合并联合训练，提升注塑件尺寸预测精度. 在注塑成型数据集上进行模型验证，预测尺寸平均均方误差为4.7×10^?4 mm²，最小误差波动为10^?5 mm²量级，模型具有较高的预测精度和稳定性.

关键词： 注塑成型; 深度学习; 双向门控循环单元网络 (BiGRU); 多特征融合; 尺寸预测

Abstract:

A multi-feature fusion injection molded part dimension prediction method based on double-layer bidirectional gated cyclic unit network (DL-BiGRU) was proposed, to fully explore the high-frequency time series features inside the mold cavity and the state features of the injection molding machine in the injection molding process. The relationship between the high-frequency time series features obtained from sensors inside the mold cavity and the dimensions of injection molded parts was analyzed. A DL-BiGRU network was utilized to automatically extract the time series features from the high-frequency data, representing the dynamic characteristics of the molding process of the injection molded parts. The instantaneous features of the injection molding process can be represented, by sampling the high-frequency time series data inside the mold cavity and flattening it. An end-to-end deep learning multi-feature fusion framework was constructed by integrating the time series features, instantaneous features, and molding machine state features. The above three characteristics were fused and jointly trained to improve the dimension prediction accuracy of injection molded parts. The model was verified on the data set of injection molding, and the results showed that the average mean square error of the predicted dimension was 4.7×10^?4 mm², and the minimum error fluctuation was on the order of 10^?5 mm². The model has high prediction accuracy and stability.

Key words: injection molding deep learning bidirectional gated cyclic unit network (BiGRU) multi-feature fusion dimension prediction

收稿日期: 2023-03-28 出版日期: 2024-03-05

CLC:

TP 181

基金资助: 国家自然科学基金资助项目(71671097); 国家重点研发计划资助项目(2019YFB1707101, 2019YFB1707103); 浙江省省属高校基本科研业务费资助项目(SJLZ2023001); 浙江省公益技术应用研究计划资助项目(LGG20E050010).

通讯作者: 余军合 E-mail: qianqingjie@foxmail.com;yujunhe@nbu.edu.cn

作者简介: 钱庆杰（1995—），男，硕士生，从事工业大数据建模与分析研究. orcid.org/0009-0007-0509-3727. E-mail：qianqingjie@foxmail.com

	服务
	把本文推荐给朋友
	加入引用管理器
	E-mail Alert
	作者相关文章
	钱庆杰
	余军合
	战洪飞
	王瑞
	胡健

引用本文:

钱庆杰,余军合,战洪飞,王瑞,胡健. 基于DL-BiGRU多特征融合的注塑件尺寸预测方法[J]. 浙江大学学报(工学版), 2024, 58(3): 646-654.

Qingjie QIAN,Junhe YU,Hongfei ZHAN,Rui WANG,Jian HU. Dimension prediction method of injection molded parts based on multi-feature fusion of DL-BiGRU. Journal of ZheJiang University (Engineering Science), 2024, 58(3): 646-654.

链接本文:

https://www.zjujournals.com/eng/CN/10.3785/j.issn.1008-973X.2024.03.021 或 https://www.zjujournals.com/eng/CN/Y2024/V58/I3/646

图 1 注塑加工过程数据采集的主要传感器位置分布图

表 1 注塑件尺寸范围及公差

表 2 传感器高频字段对应的特征

表 3 成型机状态字段对应的特征

图 2 注塑件3个模次的关键传感器高频时序数据曲线图

图 3 DL-BiGRU结构

图 4 基于DL-BiGRU多特征融合的注塑件尺寸预测模型

表 4 实验模型参数设置

表 5 不同消融模型的预测误差

图 5 不同消融模型的误差箱型图

表 6 常用尺寸预测模型和DL-BiGRU模型的预测误差对比

图 6 注塑成型数据集中DL-BiGRU模型的预测值与真实值效果对比图

1	CHEN J Y, TSENG C C, HUANG M S. Quality indexes design for online monitoring polymer injection molding [EB/OL]. (2019−08−01). https://doi.org/10.1155/2019/3720127.
2	CHAREST M, FINN R, DUBAY R. Integration of artificial intelligence in an injection molding process for on-line process parameter adjustment [C]// 2018 Annual IEEE International Systems Conference . Vancouver: IEEE, 2018: 1−6.
3	OGORODNYK O, LYNGSTAD O V, LARSEN M, et al. Application of machine learning methods for prediction of parts quality in thermoplastics injection molding [C]// Advanced Manufacturing and Automation VIII . Singapore: Springer, 2019: 237−244.
4	OGORODNYK O, LYNGSTAD O V, LARSEN M, et al. Prediction of width and thickness of injection molded parts using machine learning methods [M]// KISHITA Y, MATSUMOTO M, INOUE M, et al. EcoDesign and sustainability I: products, services, and business models . Singapore: Springer, 2021: 455−469.
5	YIN F, MAO H, HUA L, et al Back propagation neural network modeling for warpage prediction and optimization of plastic products during injection molding[J]. Materials and design, 2011, 32 (4): 1844- 1850 doi: 10.1016/j.matdes.2010.12.022
6	GÜLÇÜR M, WHITESIDE B A study of micromanufacturing process fingerprints in micro-injection moulding for machine learning and Industry 4.0 applications[J]. The International Journal of Advanced Manufacturing Technology, 2021, 115 (5/6): 1943- 1954
7	宋建, 陈广森, 陈敬福基于特征选择和贝叶斯优化 LightGBM 的注塑制品尺寸预测[J]. 工程塑料应用, 2021, 49 (8): 54- 60 SONG Jian, CHEN Guangsen, CHEN Jingfu Size prediction of injection molded products based on feature selection and bayesian optimized LightGBM[J]. Engineering Plastic Application, 2021, 49 (8): 54- 60 doi: 10.3969/j.issn.1001-3539.2021.08.010
8	宋建, 王文龙, 李东, 等基于Stacking集成学习的注塑件尺寸预测方法[J]. 华南理工大学学报:自然科学版, 2022, 50 (6): 19- 26 SONG Jian, WANG Wenlong, LI Dong, et al Injection molding part size prediction method based on Stacking ensemble learning[J]. Journal of South China University of Technology: Natural Science Edition, 2022, 50 (6): 19- 26
9	NIAN S C, FANG Y C, HUANG M S In-mold and machine sensing and feature extraction for optimized IC-tray manufacturing[J]. Polymers, 2019, 11 (8): 1348 doi: 10.3390/polym11081348
10	HUANG M S Cavity pressure based grey prediction of the filling-to-packing switchover point for injection molding[J]. Journal of Materials Processing Technology, 2007, 183 (2/3): 419- 424
11	FARAHANI S, BROWM N, LOFTIS J, et al Evaluation of in-mold sensors and machine data towards enhancing product quality and process monitoring via Industry 4.0[J]. The International Journal of Advanced Manufacturing Technology, 2019, 105: 1371- 1389 doi: 10.1007/s00170-019-04323-8
12	HOTAIT H, CHIEMENTIN X, MOUCHAWEH M S, et al Monitoring of ball bearing based on improved real-time OPTICS clustering[J]. Journal of Signal Processing Systems, 2021, 93: 221- 237 doi: 10.1007/s11265-020-01571-w
13	CHEN Z, WU M, ZHAO R, et al Machine remaining useful life prediction via an attention-based deep learning approach[J]. IEEE Transactions on Industrial Electronics, 2021, 68 (3): 2521- 2531 doi: 10.1109/TIE.2020.2972443
14	FENG T T, GUO L, GAO H, et al A new time-space attention mechanism driven multi-feature fusion method for tool wear monitoring[J]. The International Journal of Advanced Manufacturing Technology, 2022, 120: 5633- 5648 doi: 10.1007/s00170-022-09032-3
15	WICK C, EHRIG F, SCHUSTER G. Data driven injection moulding [C]// Advances in Polymer Processing 2020: Proceedings of the International Symposium on Plastics Technology . Heidelberg: Springer, 2020: 128−136.
16	陈金伟, 林旭光注塑成型模腔压力监测技术及其应用[J]. 塑料科技, 2011, 39 (8): 59- 62 CHEN Jinwei, LIN Xuguang Cavity pressure measurement technology during injection molding process and its applications[J]. Plastic Technology, 2011, 39 (8): 59- 62 doi: 10.3969/j.issn.1005-3360.2011.08.006
17	CHO K, VAN MERRIËNBOER B, GULCEHRE C, et al. Learning phrase representations using RNN encoder–decoder for statistical machine translation [C]// Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing . Doha: Association for Computational Linguistics, 2014: 1724−1734.
18	SHE D, JIA M A BiGRU method for remaining useful life prediction of machinery[J]. Measurement, 2021, 167: 108277 doi: 10.1016/j.measurement.2020.108277
19	HAN W, NAN L, SU M, et al Research on the prediction method of centrifugal pump performance based on a double hidden layer BP neural network[J]. Energies, 2019, 12 (14): 2709 doi: 10.3390/en12142709
20	于志刚, 辛小军, 范远洪, 等. 基于双层GRU神经网络的油田注水预测方法研究 [J]. 化工自动化及仪表, 2022, 49(01): 41-46. YU Zhigang, XIN Xiaojun, FAN Yuanhong, et al. Research on oilfield water injection prediction method based on double hidden layer GRU model [J]. Chemical Automation and Instrumentation , 2022, 49(1): 41−46.

[1]	宋明俊,严文,邓益昭,张俊然,涂海燕. 轻量化机器人抓取位姿实时检测算法[J]. 浙江大学学报(工学版), 2024, 58(3): 599-610.
[2]	姚鑫骅,于涛,封森文,马梓健,栾丛丛,沈洪垚. 基于图神经网络的零件机加工特征识别方法[J]. 浙江大学学报(工学版), 2024, 58(2): 349-359.
[3]	孙雪菲,张瑞峰,关欣,李锵. 强化先验骨架结构的轻量型高效人体姿态估计[J]. 浙江大学学报(工学版), 2024, 58(1): 50-60.
[4]	郑超昊,尹志伟,曾钢锋,许月萍,周鹏,刘莉. 基于时空深度学习模型的数值降水预报后处理[J]. 浙江大学学报(工学版), 2023, 57(9): 1756-1765.
[5]	杨哲,葛洪伟,李婷. 特征融合与分发的多专家并行推荐算法框架[J]. 浙江大学学报(工学版), 2023, 57(7): 1317-1325.
[6]	李云红,段姣姣,苏雪平,张蕾涛,于惠康,刘杏瑞. 基于改进生成对抗网络的书法字生成算法[J]. 浙江大学学报(工学版), 2023, 57(7): 1326-1334.
[7]	权巍,蔡永青,王超,宋佳,孙鸿凯,李林轩. 基于3D-ResNet双流网络的VR病评估模型[J]. 浙江大学学报(工学版), 2023, 57(7): 1345-1353.
[8]	周欣磊,顾海挺,刘晶,许月萍,耿芳,王冲. 基于集成学习与深度学习的日供水量预测方法[J]. 浙江大学学报(工学版), 2023, 57(6): 1120-1127.
[9]	刘沛丰,钱璐,赵兴炜,陶波. 航空装配领域中命名实体识别的持续学习框架[J]. 浙江大学学报(工学版), 2023, 57(6): 1186-1194.
[10]	赵嘉墀,王天琪,曾丽芳,邵雪明. 基于GRU的扑翼非定常气动特性快速预测[J]. 浙江大学学报(工学版), 2023, 57(6): 1251-1256.
[11]	曹晓璐,卢富男,朱翔,翁立波,卢书芳,高飞. 基于草图的兼容性服装生成方法[J]. 浙江大学学报(工学版), 2023, 57(5): 939-947.
[12]	苏育挺,陆荣烜,张为. 基于注意力和自适应权重的车辆重识别算法[J]. 浙江大学学报(工学版), 2023, 57(4): 712-718.
[13]	马庆禄,鲁佳萍,唐小垚,段学锋. 改进YOLOv5s的公路隧道烟火检测方法[J]. 浙江大学学报(工学版), 2023, 57(4): 784-794.
[14]	曾耀,高法钦. 基于改进YOLOv5的电子元件表面缺陷检测算法[J]. 浙江大学学报(工学版), 2023, 57(3): 455-465.
[15]	兰欢,余建波. 基于深度学习三维成型的钢板表面缺陷检测[J]. 浙江大学学报(工学版), 2023, 57(3): 466-476.

Viewed

Full text

Abstract

Cited

Shared

Discussed