基于超级基神经网络的自适应反演非奇异滑模纱线恒张力控制

RichHTML

PDF (PC)

基于超级基神经网络的自适应反演非奇异滑模纱线恒张力控制

Hyper basis function-based adaptive inverse non-singular method for constant-tension yarn transport

基于超级基神经网络的自适应反演非奇异滑模纱线恒张力控制

Hyper basis function-based adaptive inverse non-singular method for constant-tension yarn transport

摘要/Abstract

图/表 10

参考文献 18

Metrics

摘要/Abstract

引用本文

使用本文

图/表 10

参考文献 18

相关文章 15

Metrics

本文评价

推荐阅读 0

doi:10.13475/j.fzxb.20240904301

纺织学报 ›› 2025, Vol. 46 ›› Issue (02): 92-99.doi: 10.13475/j.fzxb.20240904301

王罗俊¹^,², 彭来湖¹(), 熊叙一¹, 李杨², 胡旭东¹

1.浙江理工大学浙江省现代纺织装备技术重点实验室, 浙江杭州 310018
2.浙江机电职业技术大学自动化学院, 浙江杭州 310053

收稿日期:2024-09-29 修回日期:2024-10-23 出版日期:2025-02-15 发布日期:2025-03-04
通讯作者: 彭来湖(1980—),男,教授,博士。主要研究方向为针织装备控制技术。E-mail:laihup@zstu.edu.cn。
作者简介:王罗俊(1993—),男,讲师,博士生。主要研究方向为针织装备控制技术。
第一联系人：
说明:本文入围中国纺织工程学会第25届陈维稷论文卓越行动计划
基金资助:
国家重点研发计划项目(SQ2023YFB3200093);浙江省“高层次人才特殊支持计划”科技创新领军人才项目(2023R5212);浙江机电职业技术大学科教融合重点培育项目(A-0271-24-209);浙江省教育厅一般科研项目(Y202456420)

WANG Luojun¹^,², PENG Laihu¹(), XIONG Xuyi¹, LI Yang², HU Xudong¹

1. Key Laboratory of Modern Textile Machinery & Technology of Zhejiang Province, Zhejiang Sci-Tech University, Hangzhou, Zhejiang 310018, China
2. School of Automation, Zhejiang Polytechnic University of Mechanical and Electrical Engineering, Hangzhou, Zhejiang 310053, China

Received:2024-09-29 Revised:2024-10-23 Published:2025-02-15 Online:2025-03-04

摘要：

为解决针织圆机高速工作时纱线张力波动较大问题,提出了一种基于超级基(HBF)神经网络区间观测器的反演非奇异滑模纱线恒张力控制方法。通过构建运动纱线系统的数学模型,运用神经网络逼近系统参数(输纱器与编织机构转动惯量)变动所导致的不确定性响应,将HBF神经网络与区间观测器相结合设计了一个区间状态观测器,估算出系统转速及纱线张力的边界范围,提高了状态识别的准确性。基于纱线张力估算值,构建反演非奇异终极滑模控制器,确保了张力跟踪误差能够在短时间内迅速收敛,从而增强了系统的鲁棒性与动态响应能力。仿真和实验结果表明:所提控制方法成功地使运动纱线张力在1.6 s内达到并维持在预设值,调节时间相较于标准滑模控制及现有文献中的滑模控制器分别缩短了57%和33%,验证了该控制算法的高效性与可靠性。

关键词: 纱线张力, 超级基神经网络, 状态观测器, 张力误差, 滑模控制器, 针织圆机

Abstract:

Objective In high-speed and precision knitting process, the complex dynamic behavior of yarn transmission not only affects the accuracy of tension control, but also increases the complexity and maintenance cost of the system. It is hence necessary to explore new control methods to improve the accuracy and reliability of yarn tension stability control. Sensorless tension control method reduces the dependence on sensors. By optimizing the structure and material of the yarn transmission mechanism, the influence of adverse factors such as vibration and friction is reduced, and the production efficiency and product quality of the circular machine are improved.

Method The yarn motion during knitting was decoupled into two independent systems using the inversion method, and an inverse non-singular terminal sliding mode controller was designed to improve the sliding mode surface to make the yarn real-time tension error converge quickly in a short time. The hyper basis function (HBF) neural network was introduced into the interval state observer of the yarn transmission system, which was close to the random response caused by the changes of parameters such as weft storage radius and the inertia of the knitting area.

Results The designed HBF neural network interval observer was used to estimate the boundary value of the moving yarn system. After the operation of the three controllers, the controller designed in this research was shown to stabilize the tension in 1.6, which is significantly better than the 3.5 s of the conventional sliding mode and the 2.4 s described in related literature, and the adjustment time was reduced by 57% and 33% respectively. The experimental results showed that the sliding mode controller designed in this paper has faster response and higher tracking accuracy, which is significantly better than the other two controllers. The traditional proportional-integral-differential (PID) controller performed the worst for the yarn relaxation problem when the moving yarn system is started, while the improved sliding mode controller can stabilize the winding speed faster. In addition, the terminal sliding mode controller designed in this paper can quickly restore the tension stability after the random disturbance is added in the 8^th s of the system movement, showing excellent robust performance. The sliding mode controller quickly converges the tension error to zero within 1.6 s after the start-up of the winding system. Compared with the other three controllers, the sliding mode controller has the optimal response speed and adjustment ability, ensuring that the tension control system can recover to steady state operation in a short time. It obviously improves the steady-state operation ability and dynamic adjustment ability of the yarn system, so as to realize the constant tension control of the yarn. Through experiments, it can be verified that the method in this paper can quickly and accurately follow the change of the target tension, and has low sensitivity to external interference. Even in the face of the sudden change of the tension setting, it can effectively inhibit the overshoot, show stronger stability and response speed, and has better robustness, faster response speed and higher control accuracy in the actual production environment, which is more in line with the production requirements.

Conclusion The relationship between yarn tension and motion speed is established by modeling the motion yarn system of yarn feeder and loop forming mechanism. The neural network technology is used to approximate the influence of unknown time-varying parameters, and an interval observer is constructed based on it to realize the effective observation of key state variables. The traditional nonsingular fast terminal sliding mode controller is improved. By designing a new sliding mode surface function, not only the finite time convergence of tracking error is guaranteed, but also the convergence rate is accelerated. Combined with the inversion control algorithm, the robustness and stability of the system are significantly enhanced. Simulation and experimental results show that the proposed RBF neural network interval observer can accurately track the system state and improve the control accuracy. Compared with the traditional method, the improved sliding mode controller shows faster error convergence speed and higher response efficiency.

Key words: yarn tension, hyper basis function neural network, state observer, tension error, sliding mode controller, circular knitting machine

中图分类号:

TS181.8

王罗俊, 彭来湖, 熊叙一, 李杨, 胡旭东. 基于超级基神经网络的自适应反演非奇异滑模纱线恒张力控制[J]. 纺织学报, 2025, 46(02): 92-99.

WANG Luojun, PENG Laihu, XIONG Xuyi, LI Yang, HU Xudong. Hyper basis function-based adaptive inverse non-singular method for constant-tension yarn transport[J]. Journal of Textile Research, 2025, 46(02): 92-99.

导出引用管理器 EndNote|Reference Manager|ProCite|BibTeX|RefWorks

链接本文: http://www.fzxb.org.cn/CN/10.13475/j.fzxb.20240904301

http://www.fzxb.org.cn/CN/Y2025/V46/I02/92

图1

图2

表1

图3

图4

图5

图6

图7

图8

图9

[1]	李杨, 彭来湖, 刘建廷, 等. 基于横向振动频率的轴向运动纱线张力测量[J]. 纺织学报, 2023, 44(6): 72-77.
	LI Yang, PENG Laihu, LIU Jianting, et al. Axial yarn tension measurement based on transverse vibration frequency[J]. Journal of Textile Research, 2023, 44(6): 72-77.
[2]	章钰娟, 彭来湖, 徐郁山, 等. 非接触式纱线状态检测技术研究[J]. 现代纺织技术, 2021(10): 1-10.
	ZHANG Yujuan, PENG Laihu, XU Yushan, et al. Research on non-contact yarn state detection techno-logy[J]. Advanced Textile Technology, 2021(10): 1-10.
[3]	成玲, 梁银铮. 纱线的动态力学性能[J]. 纺织学报, 2006, 27 (8): 19-21.
	CHENG Ling, LIANG Yinzheng. The dynamic mechanics performance of the yarn [J]. Journal of Textile Research, 2006, 27(8): 19-21.
[4]	戴宁, 胡旭东, 彭来湖. 针织大圆机运动实时控制技术[J]. 纺织学报, 2019, 40(12): 134-139. doi: 10.13475/j.fzxb.20180705606
	DAI Ning, HU Xudong, PENG Laihu. Real time motion control technology circular knitting machine[J]. Journal of Textile Research, 2019, 40(12): 134-139. doi: 10.13475/j.fzxb.20180705606
[5]	周俊儒, 林光兴, 张彦, 等. 圆纬机智能送纱系统开发与技术优势分析[J]. 纺织导报, 2022(6): 80-83.
	ZHOU Junru, LIN Guangxing, ZHANG Yan, et al. Development and technical advantage analysis of intelligent yarn feeding system for circular weft knitting machine[J]. China Textile Leader, 2022(6): 80-83.
[6]	周佳超. 高弹细纱输送动态控制技术[D]. 杭州: 浙江理工大学, 2021:14-18.
	ZHOU Jiachao. Dynamic control technology of high elastic yarn conveying[D]. Hangzhou: Zhejiang Sci-Tech University, 2021:14-18.
[7]	李雪娇. 短纤纱高速经编动态张力调控系统研究[D]. 无锡: 江南大学, 2022:14-16.
	LI Xuejiao. Research on dynamic tension control system for high speed warp knitting of staple fiber yarn[D]. Wuxi: Jiangnan University, 2022:14-16.
[8]	ALI M, AHMED R, AMER M. Yarn tension control technique for improving polyester soft winding process[J]. Scientific Reports, 2021, 11(1):1060-1075. doi: 10.1038/s41598-020-79928-1 pmid: 33441729
[9]	ANDRE, MATTHES, CHOKRI, et al. Analysis and simulation of yarn feeding at high dynamic circular knitting machines[J]. Melliand Interna-tional(Worldwide Textile Journal), 2010, 16(5/6):220-222.
[10]	李鹏飞, 高文根, 张港, 等. 基于改进SMO的无传感器IPMSM 转子位置估计[J]. 电子测量与仪器学报, 2021, 35(9):65-72.
	LI Pengfei, GAO Wengen, ZHANG Gang, et al. Sensorless IPMSM rotor position estimation based on improved smo[J]. Journal of Electronic Measurement and Instrumentation, 2021, 35 (9): 65-72.
[11]	易泽仁, 谢巍, 刘龙文, 等. 一类非线性系统的神经网络自适应区间观测器设计[J]. 控制理论与应用, 2023, 40(10):1730-1736.
	YI Zeren, XIE Wei, LIU Longwen, et al. Design of neural network adaptive interval observer for a class of nonlinear systems[J]. Control Theory and Applications, 2023, 40 (10): 1730-1736.
[12]	胡军剑. 大惯性电液张力模拟绞车系统的动特性及张力控制研究[D]. 杭州: 浙江大学, 2020:12-13.
	HU Junjian. Research on dynamic characteristics and tension control of large inertia electro-hydraulic tension simulation winch system[D]. Hangzhou: Zhejiang University, 2020:12-13.
[13]	楚晓艳, 年晓红, 刘静静. 基于鲁棒滑模观测器的多电机卷绕系统故障检测和隔离[J]. 控制理论与应用, 2018, 35(6):795-804.
	CHU Xiaoyan, NIAN Xiaohong, LIU Jingjing. Fault detection and isolation of multi motor winding system based on robust sliding mode observer[J]. Control Theory and Application, 2018, 35(6): 795-804.
[14]	郁明, 李旺林, 蓝盾. 基于优化自适应阈值的非线性机电系统传感器故障检测和主动容错控制[J]. 仪器仪表学报, 2022, 43(4):26-37.
	YU Ming, LI Wanglin, LAN Dun. Sensor fault detection and active fault tolerant control of nonlinear electromechanical systems based on optimized adaptive threshold[J]. Journal of Instrumentation, 2022, 43(4): 26-37.
[15]	何忠伟. 区间观测器及其控制系统研究[D]. 广州: 华南理工大学, 2016:15.
	HE Zhongwei. Research on interval observer and its control system[D]. Guangzhou: South China University of Technology, 2016:15.
[16]	付光杰, 潘海龙. 基于神经网络观测器的 PMSM无传感器矢量控制[J]. 国外电子测量技术, 2018, 37(5): 98-101.
	FU Guangjie, PAN Hailong. Sensorless vector control of PMSM Based on neural network observer[J]. Foreign Electronic Measurement Technology, 2018, 37(5): 98-101.
[17]	武晨, 谷松原, 房圣超. HBF神经网络的一种结构自适应在线学习算法[J]. 中国电子科学研究院学报, 2021, 16(5):486-495.
	WU Chen, GU Songyuan, FANG Shengchao. A structure adaptive online learning algorithm for HBF neural network[J]. Journal of China Academy of Electronic Sciences, 2021, 16 (5): 486-495.
[18]	周锋, 郑丽霞, 沈烨超, 等. 基于自适应反演的超细金刚线张力控制系统[J]. 中国机械工程, 2021, 32(24): 2995-3000.
	ZHOU Feng, ZHENG Lixia, SHEN Yechao, et al. Tension control system of superfine diamond wire based on adaptive inversion[J]. China Mechanical Engineering, 2021, 32 (24): 2995-3000.

Viewed

Full text

From	Others	local

Times	10	10
Rate	50%	50%

Abstract

Just accepted	Online first	Issue

0	0	52

From	Others	local

Times	44	8
Rate	85%	15%

Cited

Web of Science	Crossref	ScienceDirect	Search for Citations in Google Scholar >>


This page requires you have already subscribed to WoS.

Shared

Discussed

参数	单位	参数取值
E	cN/m²	1.2×10⁴
$ρ$	tex	27.4
L	m	1.2
$ε$	kg/m²	1.0×10^-3
J	kg/m²	2×10^-4
$R 0$	m	0.5
$F 0$	cN	30
$V 0$	m/s	0.6

Just accepted

Online first

Just accepted

Online first

[1]	蒋静, 彭来湖, 史伟民, 袁豪伟. 基于奇异值分解算法的非接触纱线张力测量[J]. 纺织学报, 2024, 45(09): 204-211.
[2]	彭来湖, 侯良美, 齐育宝, 汝欣, 刘建廷. 基于自适应Loess的纱线张力信号处理方法[J]. 纺织学报, 2024, 45(02): 246-254.
[3]	孙磊, 屠佳佳, 毛慧敏, 王俊茹, 史伟民. 针织智能车间自动换筒任务调度技术[J]. 纺织学报, 2023, 44(12): 189-196.
[4]	彭来湖, 刘建廷, 李杨, 齐育宝, 李建强, 茅木泉. 基于卷积滤波的接触式纱线张力测量方法[J]. 纺织学报, 2023, 44(08): 73-80.
[5]	李杨, 彭来湖, 刘建廷, 胡旭东, 郑秋扬. 基于横向振动频率的轴向运动纱线张力测量[J]. 纺织学报, 2023, 44(06): 72-77.
[6]	纪越, 潘东, 马杰东, 宋丽梅, 董九志. 基于机器视觉的弦振动纱线张力非接触检测系统[J]. 纺织学报, 2023, 44(05): 198-204.
[7]	应志平, 王伟青, 吴震宇, 胡旭东. 三维正交机织复合材料的冲后压缩性能[J]. 纺织学报, 2023, 44(01): 129-135.
[8]	彭来湖, 章钰娟, 吕永法, 戴宁, 李建强. 纬编针织纱线输送状态检测方法及其动态特性[J]. 纺织学报, 2022, 43(12): 167-172.
[9]	郭敏, 高卫东, 朱博, 刘建立, 郭明瑞. 模拟织造状态下的浆纱耐磨性能测试方法[J]. 纺织学报, 2021, 42(11): 46-50.
[10]	孙帅, 缪旭红, 张琦, 王瑾. 高速经编机上纱线张力的波动规律[J]. 纺织学报, 2020, 41(03): 51-55.
[11]	徐云龙, 夏风林. 双针床经编机梳栉摆动对瞬时需纱量和纱线张力的影响[J]. 纺织学报, 2019, 40(06): 106-110.
[12]	孙帅缪旭红张灵婕胡瑜. 经编纱线张力补偿装置的工作机制[J]. 纺织学报, 2018, 39(11): 140-144.
[13]	胡瑜刘行缪旭红. 经编纱线动态张力评价指标[J]. 纺织学报, 2018, 39(02): 68-72.
[14]	扈昕瞳张玉井孟婥孙以泽. 编织锭子放线速度对纱线张力调控的建模与影响[J]. 纺织学报, 2017, 38(06): 111-117.
[15]	夏胜华孙以泽孟婥任国斌 . 簇绒地毯织机提花装置的绕纱动态张力分析[J]. 纺织学报, 2015, 36(07): 136-141.