纺织学报 ›› 2022, Vol. 43 ›› Issue (12): 75-81.doi: 10.13475/j.fzxb.20210607107

• 纺织工程 • 上一篇    下一篇

基于图像模拟和支持向量机的复合非织造布孔隙尺寸预测

金关秀1(), 祝成炎2   

  1. 1.浙江工业职业技术学院, 浙江 绍兴 312000
    2.浙江理工大学 纺织科学与工程学院(国际丝绸学院), 浙江 杭州 310018
  • 收稿日期:2021-06-24 修回日期:2022-09-04 出版日期:2022-12-15 发布日期:2023-01-06
  • 作者简介:金关秀(1962—),男,教授,博士。主要研究方向为非织造材料的结构与性能。E-mail: ctljgx@163.com。

Prediction of pore dimension in composite nonwovens based on image simulation and support vector machine

JIN Guanxiu1(), ZHU Chengyan2   

  1. 1. Zhejiang Industry Polytechnic College, Shaoxing, Zhejiang 312000, China
    2. College of Textile Science and Engineering (International Silk Institute), Zhejiang Sci-Tech University, Hangzhou, Zhejiang 310018, China
  • Received:2021-06-24 Revised:2022-09-04 Published:2022-12-15 Online:2023-01-06

RichHTML

8

PDF (PC)

64

摘要：

为探究非织造布复合前后孔隙尺寸的变化规律,以纤维根数、纤维平均直径、纤维直径变异系数为变量生成数字图像(复合单元)对非织造纤维网进行模拟,并采用将2个复合单元叠加的手法来模拟复合后的非织造纤维网。用2个复合单元的孔径差异百分比(Er)和孔径变异系数的平均值(MV)表征复合前孔隙尺寸;孔径复合指数(IP) 和孔径变异系数复合指数(IV)用于表征复合后孔隙尺寸。结果显示非织造布复合前后的孔隙尺寸呈复杂的非线性关系。将Er和MV作为支持向量机模型的输入分别对IP和IV进行预测,其绝对百分比误差分别为1.84%和1.92%,误差很小。验证实验的结果进一步印证支持向量机模型具有很高的预测准确度。

关键词: 复合非织造布, 孔径, 孔径变异系数, 数字图像模拟, 支持向量机

Abstract:

In order to study the influence of nonwoven compounding on pore dimension in composite nonwovens between before and after compounded, the nonwoven webs was modeled based on the use of digital images taking into considerations of fiber number, the average fiber diameter and the variation coefficient of fiber diameter served as the model variables. Thus the image model of composite nonwoven was then obtained by superimposing two compound elements. The pore size percentage difference (Er) and the average pore size variation coefficient (Mv) of two compound elements were used to characterize the pore dimension of nonwoven before compounded. Furthermore, the pore size composite index (IP) and pore size variation coefficient composite index (IV) were used to characterize the pore dimension after compounding. The test results show that there is a complex non-linear relationship on pore dimension between before and after compounding. Er and Mv were used as the inputs of support vector machine model to predict IP and IV, respectively. The results indicate that the mean absolute percentage errors of the above two predictions are only 1.84% and 1.92%, respectively. The verification experiment result further confirms the high prediction accuracy of the support vector machine model.

Key words: composite nonwoven, pore size, pore size variation coefficient, digital image simulation, support vector machine

中图分类号: 

  • TS171

表1

复合单元建模方案"

复合单元编号 n d/μm Vd/%
A 24 10 10
B 24 10 30
C 24 20 10
D 24 20 30
E 36 10 10
F 36 10 30
G 36 20 10
H 36 20 30

图1

复合单元数字图像模拟结果"

表2

复合单元孔径及其变异系数的测定结果"

复合单元编号 孔径/μm 孔径变异系数/%
A 32.9 85.39
B 28.2 45.31
C 34.9 96.53
D 31.7 52.30
E 22.0 74.56
F 24.5 39.75
G 21.6 78.66
H 21.3 49.04

表3

模拟复合方案"

复合图
像编号		 复合
单元1		 复合
单元2		 复合图
像编号		 复合
单元1		 复合
单元2
1# A B 15# C E
2# A C 16# C F
3# A D 17# C G
4# A E 18# C H
5# A F 19# D E
6# A G 20# D F
7# A H 21# D G
8# B C 22# D H
9# B D 23# E F
10# B E 24# E G
11# B F 25# E H
12# B G 26# F G
13# B H 27# F H
14# C D 28# G H

图2

模拟复合结果"

表4

复合图像孔径及其变异系数测定结果"

复合图
像编号		 孔径/
μm		 孔径变异
系数/%		 复合图
像编号		 孔径/
μm		 孔径变异
系数/%
1# 12.9 140.65 15# 11.1 144.10
2# 12.7 196.72 16# 12.9 135.21
3# 13.1 163.23 17# 11.2 143.58
4# 10.2 132.76 18# 13.5 112.80
5# 14.2 116.82 19# 13.2 91.02
6# 10.1 131.49 20# 15.8 46.09
7# 13.0 97.57 21# 12.7 100.23
8# 12.5 166.76 22# 14.4 41.56
9# 15.3 67.15 23# 10.9 109.64
10# 12.2 94.80 24# 8.6 158.53
11# 15.5 50.27 25# 9.6 133.69
12# 12.1 101.79 26# 10.4 113.51
13# 13.8 42.75 27# 12.8 55.69
14# 12.8 187.31 28# 9.3 147.38

表5

图像复合前后的孔隙尺寸"

复合图像编号 Er/% MV/% IP IV 复合图像编号 Er/% MV/% IP IV
1# 16.67 65.35 0.457 4 1.647 1 15# 58.64 85.55 0.504 5 1.492 8
2# 6.08 90.96 0.386 0 2.037 9 16# 42.45 68.14 0.526 5 1.400 7
3# 3.79 68.85 0.413 2 1.911 6 17# 61.57 87.60 0.518 5 1.487 4
4# 49.55 79.98 0.463 6 1.554 7 18# 63.85 72.79 0.633 8 1.168 5
5# 34.29 62.57 0.579 6 1.368 1 19# 44.09 63.43 0.600 0 1.220 8
6# 52.31 82.03 0.467 6 1.539 9 20# 29.39 46.03 0.644 9 0.881 3
7# 54.46 67.22 0.610 3 1.142 6 21# 46.76 65.48 0.588 0 1.274 2
8# 23.76 70.92 0.443 3 1.727 5 22# 48.83 50.67 0.676 1 0.884 6
9# 12.41 48.81 0.542 6 1.283 9 23# 11.36 57.16 0.495 5 1.470 5
10# 28.18 59.94 0.554 5 1.271 5 24# 1.85 76.61 0.398 1 2.015 4
11# 15.10 42.53 0.632 7 1.109 5 25# 3.29 61.80 0.450 7 1.793 1
12# 30.56 61.99 0.560 2 1.294 1 26# 13.43 59.21 0.481 5 1.443 0
13# 32.39 47.18 0.647 9 0.871 7 27# 15.02 44.40 0.600 9 1.135 6
14# 10.09 74.42 0.403 8 1.940 4 28# 1.41 63.85 0.436 6 1.873 6

表6

SVM模型结构参数的优化结果与预测准确度"

预测参数 优化后的结构参数 U/%
W FP ε
IP 67 2 600 0.001 1.84
IV 69 2 800 0.002 1.92

表7

非织造布复合方案"

试样编号 复(叠)合单元
T1 S1+M1
T2 S1+M2
T3 S2+M1
T4 S2+M2

表8

非织造布复(叠)合前后孔径及其变异系数测试结果"

试样编号 孔径/
μm		 孔径变异系数/
%
复(叠)合前 S1 26.5 105.83
S2 30.9 76.40
M1 19.8 16.32
M2 14.1 25.81
复(叠)合后 T1 10.3 173.84
T2 9.7 160.12
T3 14.2 69.17
T4 12.3 66.10

表9

非织造布复(叠)合前后的孔隙尺寸"

试样编号 ErT/% MVT/% IPT IVT
T1 12.39 63.60 0.472 5 1.642 6
T2 33.88 67.82 0.530 1 1.513 0
T3 33.49 48.89 0.651 4 0.905 4
T4 59.02 53.11 0.672 1 0.865 2

表10

验证实验结果"

试样编号 孔径复合指数 孔径变异系数复合指数
T1 3.72 0.39
T2 3.73 1.51
T3 0.85 0.49
T4 1.83 4.67
[1] IRZMANSKA E, JURCZYK K M, BIL M, et al. Experimental investigation of the properties of laminated nonwovens used for packaging of powders in mineral warmers[J]. Autex Research Journal, 2021, 21(2): 163-171.
doi: 10.2478/aut-2019-0078
[2] 虞胜椿, 肖春晓, 徐鹤年, 等. 亲水性纺粘-熔喷复合非织造材料的制备及性能[J]. 现代丝绸科学与技术, 2020, 35(6): 17-19.
YU Shengchun, XIAO Chunxiao, XU Henian, et al. Preparation and properties of hydrophilic spun-bonded/melt blown composite nonwovens[J]. Modern Silk Science & Technology, 2020, 35(6): 17-19.
[3] WANG H, ZHANG Y, GAO H P. Composite melt-blown nonwoven fabrics with large pore size as Li-ion battery separator[J]. International Journal of Hydrogen Energy, 2016, 41(1): 324-330.
doi: 10.1016/j.ijhydene.2015.09.130
[4] 吴芳, 王士勇, 都丽红, 等. 非织造过滤介质孔径及截留性能的研究[J]. 无机盐工业, 2020, 52(2): 38-42.
WU Fang, WANG Shiyong, DU Lihong, et al. Study on aperture performance and rejection performance of nonwoven filter media[J]. Inorganic Chemicals Industry, 2020, 52(2): 38-42.
[5] AYDILEK A H, ASCE A M, OGUZ S H. Digital image analysis to determine pore opening size distribution of nonwoven geotextiles[J]. Journal of Computing in Civil Engineering, 2002(16): 280-290.
[6] AYDILEK A H, OGUZ S H, EDIL T B. Constriction size of geotextile filters[J]. Journal of Geotechnical and Geoenvironmental Engineering, 2005(131): 28-37.
[7] SIMMONDS G E, BOMBERGER J D, BRYNER M A. Designing nonwovens to meet pore size specifica-tions[J]. Journal of Engineering Fibers and Fabrics, 2007(2): 1-15.
[8] 潘莺, 王善元. 模拟熔喷纤网的孔径分布及其对过滤性能的影响[J]. 中国纺织大学学报, 2000, 26(5): 73-77.
PAN Ying, WANG Shanyuan. Simulation of melt blown web and the effect of its pore size distribution on filtration performance[J]. Journal of China Textile University, 2000, 26(5): 73-77.
[9] 杨旭红. 非织造材料(纤维网)形态结构的表征与分形模拟[D]. 苏州: 苏州大学, 2003: 76-87.
YANG Xuhong. Expression and fractal simulation of morphologic structures of nonwovens[D]. Suzhou: Soochow University, 2003: 76-87.
[10] DIMASSI M, KOEHL L, ZENG X, et al. Pore network modelling using image processing techniques[J]. International Journal of Clothing Science and Technology, 2008, 20 (3): 137-149.
doi: 10.1108/09556220810865193
[11] SUN L, LI Y, DU H, et al. Fault diagnosis method of low noise amplifier based on support vector machine and hidden Markov model[J]. Journal of Electronic Testing, 2021(6): 1-9.
[12] 安琪, 付译鋆, 张瑜. 医用防护服用非织造材料的研究进展[J]. 纺织学报, 2020, 41(8): 188-196.
AN Qi, FU Yijun, ZHANG Yu. Research progress of nonwovens for medical protective garment[J]. Journal of Textile Research, 2020, 41(8): 188-196.
[13] MORTEZA V, ASGHAR A Y A. Modeling the spunbonded nonwoven fabric bursting strength using finite element method[J]. Journal of The Textile Institute, 2021, 112(5): 837-845.
doi: 10.1080/00405000.2020.1783061
[1] 孙晓伦, 陈利, 张一帆, 李默涵. 开孔三维机织复合材料的拉伸性能[J]. 纺织学报, 2022, 43(08): 74-79.
[2] 徐增波, 张玲, 张艳红, 陈桂清. 基于复杂网络提取和支持向量机模型分类的服装领型研究[J]. 纺织学报, 2021, 42(06): 146-152.
[3] 董甜甜, 王蕾, 高卫东. 防羽面料孔径及分布特征与透气性和防钻绒性的关系[J]. 纺织学报, 2020, 41(12): 49-53.
[4] 潘璐, 程亭亭, 徐岚. 聚己内酯/聚乙二醇大孔径纳米纤维膜的制备及其在组织工程支架中的应用[J]. 纺织学报, 2020, 41(09): 167-173.
[5] 夏海浜, 黄鸿云, 丁佐华. 基于迁移学习与支持向量机的服装舒适度评估[J]. 纺织学报, 2020, 41(06): 125-131.
[6] 龚雪, 袁理, 刘军平, 杨亚莉, 刘沐黎, 柯政涛, 鄢煜尘. 混合色彩空间与多核学习的色纺织物组织点识别[J]. 纺织学报, 2020, 41(05): 58-65.
[7] 石康君, 王静安, 高卫东. 基于傅里叶频谱特征的织物平整度客观评级[J]. 纺织学报, 2019, 40(11): 50-56.
[8] 陶开鑫, 俞成丙, 侯颀骜, 吴聪杰, 刘引烽. 基于最小二乘支持向量机的棉针织物活性染料湿蒸染色预测模型[J]. 纺织学报, 2019, 40(07): 169-173.
[9] 李佳平 沈国康 欧耀明 孟想 辛斌杰. 应用连续投影算法及最小二乘支持向量机的单组分纺织品识别[J]. 纺织学报, 2018, 39(08): 46-51.
[10] 金关秀 张毅 楼永平 祝成炎. 应用粗糙集和支持向量机的熔喷非织造布过滤性能预测[J]. 纺织学报, 2018, 39(06): 142-148.
[11] 张恒 甄琪 钱晓明 杨红英 申屠宝卿 张一风 刘让同. 聚酯/聚酰胺中空橘瓣型超细纤维非织造材料的孔径预测[J]. 纺织学报, 2018, 39(01): 56-61.
[12] 张陆佳 林兰天 陈春敏 申炎仃 高琮. 基于主成分分析的纤维拉伸断裂声发射信号识别[J]. 纺织学报, 2018, 39(01): 19-24.
[13] 张建新 张银露 胡旭东. 光谱优化处理结合多层次支持向量机的混合染液浓度检测方法[J]. 纺织学报, 2017, 38(07): 90-94.
[14] 李东 万贤福 汪军. 采用傅里叶描述子和支持向量机的服装款式识别方法[J]. 纺织学报, 2017, 38(05): 122-127.
[15] 张磊 郝露 徐山青. 废弃棉/亚麻混纺织物制备活性炭的性能[J]. 纺织学报, 2016, 37(08): 21-25.
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
No Suggested Reading articles found!
京ICP备14006648号　 京公网安备11010502044800号
版权所有 © 2021 《纺织学报》编辑部
地址： 北京市朝阳区延静里中街3号主楼6层《纺织学报》杂志社　邮政编码：100025　
电话：010-65017711,65917740　传真：010-65016539　Email:fangzhixuebao@vip.126.com
本系统由北京玛格泰克科技发展有限公司设计开发