Application and Selection of Different Clustering Algorithms in the Processing of Impact-resistant Finite Element Simulation Results

doi:10.12382/bgxb.2025.0217

Abstract

Abstract:

In order to investigate the performance of different clustering algorithms in processing the finite element resultant stress nephograms,a single-layer ballistic impact finite element model of Twaron^® plain weave fabric is established.Four different clustering algorithms,namely,k-means,Gaussian mixture model (GMM),Mean-shift,and density-based spatial clustering of applications with noise (DBSCAN),are used to cluster the stress nephograms and analyze the results comparatively by taking the resultant images of the stress distributions as an example.The results show that the Mean-shift and DBSCAN clustering algorithms are not suitable for processing a large number of finite element stress stress mapss,the k-means and GMM clustering algorithms improve the processing efficiency by 74.24 and 172.64 times compared with the traditional manual processing,and the GMM clustering algorithm produces errors when the color of the image is not clearly differentiated.The k-means clustering algorithm ensures high efficiency while keeping the error within 0.85%.Therefore,among these four algorithms,the k-means clustering algorithm is the most suitable for fast,objective and quantitative analysis of a large number of stress nephograms.Using k-means clustering algorithm,it is measured that the area of the stress interval of single-layer Twaron^® plain weave fabrics decreases with a stress area change rate of 5.61×10⁷mm²/s for 0-600MPa,and the area of the stress interval increases with a stress area change rate of 5.27×10⁷mm²/s for 600~1200MPa within 1-15μs of the impacts.

Key words: stress distribution, ballistic impact, clustering algorithm, computer vision, finite element analysis

CLC Number:

TS 941.2

TAO Yubo, LIN Xinyi, HE Jie, YUAN Zishun, XU Wang. Application and Selection of Different Clustering Algorithms in the Processing of Impact-resistant Finite Element Simulation Results[J]. Acta Armamentarii, 2025, 46(11): 250217-.

Figures/Tables 17

Fig.1 Yarn FE model

Fig.2 Bulletproof fabrics and bullet FE model

Table 1 Material parameters of finite element model

材料参数	符号	Twaron^®纱线	弹丸
密度/(g·mm^-3)	ρ	1.44×10^-3	7.800×10^-3
屈服应力/MPa	σ	3.95×10³	刚体
断裂应变/%	ε	1.00×10^-6	刚体
杨氏模量/MPa	E₁	7.20×10⁴	2.068×10⁵
	E₂	900
	E₃	900
泊松比	ν₁₂	0.01	0.300
	ν₁₃	0.01
	ν₂₃	0.01
剪切模量/MPa	G₁₂	360.00
	G₁₃	360.00
	G₂₃	360.00

Fig.3 Comparison of finite element simulation and ballistic experimental results

Fig.4 Brief classification of clustering algorithms

Fig.5 Sample image

Fig.6 Example images manually processed using ImageJ® software

Fig.7 Image processing results using Mean-Shift clustering algorithm

Fig.8 Image processing results using Mean-Shift clustering algorithm

Fig.9 Sum of error squares and the elbow plot of SSE and cluster number k

Fig.10 Image processing results using the k-means clustering algorithm

Fig.11 The processed results of the GMM clustering algorithm

Fig.12 Comparison among the processed results of two algorithms and the manual processing results

Table 2 The errors of processing the example image 1 by two algorithms

应力区间/ MPa	k-means算法		GMM算法
应力区间/ MPa	绝对误差/ mm²	相对误差/%	绝对误差/ mm²	相对误差/ %
0~300	2.12	0.17	971.41	80.18
300~600	1.83	0.12	971.73	61.25
600~900	0.62	0.35	2.65	1.49
900~1200	0.50	0.22	2.67	1.16

Table 3 The errors of prodrssing the example image 2 by two algorithms

应力区间/ MPa	k-means算法		GMM算法
应力区间/ MPa	绝对误差/ mm²	相对误差/%	绝对误差/ mm²	相对误差/ %
0~300	3.11	0.27	1 280.35	112.81
300~600	2.99	0.21	1280.33	87.95
600~900	1.39	0.85	4.81	2.93
900~1200	1.44	0.32	5.09	1.14

Fig.13 The processeed image results of the k-means clustering algorithm

Fig.14 Change in train energy in the overall model

References 29

[1]	郭金基, 梁浩云, 许兆銮, 等. 防弹模拟靶动态应力和位移的测定及分析[J]. 系统工程与电子技术, 1993, 15(10):70-74.
	GUO J J, LIANG H Y, XU Z L, et al. Determination and analysis of dynamic stress and displacement in imitative target wearing bulletproof vest[J]. Journal of Systems Engineering and Electronics, 1993, 15(10):70-74. (in Chinese)
[2]	YANG Y F, CHEN X G. Investigation of energy absorption mechanisms in a soft armor panel under ballistic impact[J]. Textile Research Journal, 2017, 87(20):2475-2486. doi: 10.1177/0040517516671129 URL
[3]	LU B Q, ZHANG J Y, ZHENG D F, et al. Theoretical analysis on carbon fiber reinforced aluminum laminate under off-center impact[J]. International Journal of Mechanical Sciences, 2023, 248:108247. doi: 10.1016/j.ijmecsci.2023.108247 URL
[4]	HE J, YUAN Z S, XU W, et al. Computer vision for enhanced quantification of FEA of ballistic impact[J]. International Journal of Mechanical Sciences, 2024, 283:109630. doi: 10.1016/j.ijmecsci.2024.109630 URL
[5]	王纪龙, 温垚珂, 刘东旭, 等. 基于键基近场动力学对防弹插板在步枪弹侵彻下的力学响应仿真[J]. 兵工学报, 2024, 45(11):4094-4105. doi: 10.12382/bgxb.2024.0275
	WANG J L, WEN Y K, LIU D X, et al. Mechanical response of body armor during rifle bullet penetration based on bond-based peridynamic method[J]. Acta Armamentarii, 2024, 45(11):4094-4105. (in Chinese) doi: 10.12382/bgxb.2024.0275
[6]	汪萌, 温垚珂, 徐诚, 等. 步枪弹侵彻背部穿硬防护人体靶标钝击效应研究[J]. 兵器装备工程学报, 2024, 45(4):119-127.
	WANG M, WEN Y K, XU C, et al. Study on the blunt impact effects of rifle bullet penetrating the back through human torso targets wear ballistic insert plate[J]. Journal of Ordnance Equipment Engineering, 2024, 45(4):119-127. (in Chinese)
[7]	YUAN Z S, WANG K C, QIU J W, et al. A numerical study on the mechanisms of Dyneema^® quasi-isotropic woven panels under ballistic impact[J]. Composite Structures, 2020, 236:111855. doi: 10.1016/j.compstruct.2020.111855 URL
[8]	LIU J T, GAO Y Z, NIU B B, et al. Biomechanical analysis of a novel clavicular hook plate for the treatment of acromioclavicular joint dislocation:a finite element analysis[J]. Computers in Biology and Medicine, 2021, 133:104379. doi: 10.1016/j.compbiomed.2021.104379 URL
[9]	YUAN Z S, MA W F, XU W, et al. A numerical study on stress wave propagation in quasi-isotropic stacks of Dyneema^® compliant composite laminates[J]. Composite Structures, 2023, 312:116869. doi: 10.1016/j.compstruct.2023.116869 URL
[10]	解亚宸, 黄广炎, 张宏, 等. UHMWPE纤维二维织物抗弹道冲击性能[J]. 兵工学报, 2022, 43(9):2152-2163.
	XIE Y C, HUANG G Y, ZHANG H, et al. Ballistic performance of two-dimensional UHMWPE fabric[J]. Acta Armamentarii, 2022, 43(9):2152-2163. (in Chinese) doi: 10.12382/bgxb.2021.0648
[11]	马王菲, 胡裕鹏, 袁子舜, 等. 层后空间对防弹织物防护性能的影响[J]. 纺织高校基础科学学报, 2023, 36(6):8-14.
	MA W F, HU Y P, YUAN Z S, et al. Effect of backspace on the protective performance of bullet-proof fabric[J]. Basic Sciences Journal of Textile Universities, 2023, 36(6):8-14. (in Chinese)
[12]	甄泓, 肖李军, 杜成鑫, 等. 在弹道冲击条件下超高分子量聚乙烯纤维复合材料板破坏模式[J]. 兵工学报, 2025, 46(7):240725. doi: 10.12382/bgxb.2024.0725
	ZHEN H, XIAO L J, DU C X, et al. Damage modes of ultra-high molecular weight polyethylene fiber composite plates under ballistic impact conditions[J]. Acta Armamentarii, 2025, 46(7):240725. (in Chinese) doi: 10.12382/bgxb.2024.0725
[13]	LEE H, CHOI C W, JIN J W, et al. Homogenization-based multiscale analysis for equivalent mechanical properties of nonwoven carbon-fiber fabric composites[J]. Journal of Mechanical Science and Technology, 2019, 33(10):4761-4770. doi: 10.1007/s12206-019-0917-6
[14]	YANG Y F. Study on ballistic performance of hybrid soft body armour[D]. Manchester, UK: The University of Manchester, 2016:94-123.
[15]	ROYLANCE D, WILDE A, TOCCI G. Ballistic impact of textile structures[J]. Textile Research Journal, 1973, 43(1):34-41. doi: 10.1177/004051757304300105 URL
[16]	ROYLANCE D, WANG S. Penetration mechanics of textile structures[M]. New York,NY,US: Elsevier, 1980:1-30.
[17]	ROYLANCE D. Stress wave propagation in fibers:Effects of crossovers[J]. Fiber Science and Technology, 1980, 13(5):385-395. doi: 10.1016/0015-0568(80)90011-1 URL
[18]	袁子舜, 陆振乾, 许玥, 等. 超高分子量聚乙烯纤维平纹织物-单向布混合堆叠板的防弹机制[J]. 复合材料学报, 2022, 39(6):2707-2715.
	YUAN Z S, LU Z Q, XU Y, et al. Ballistic mechanism of the hybrid panels with UHMWPE woven fabrics and UD laminates[J]. Acta Materiae Compositae Sinica, 2022, 39(6):2707-2715. (in Chinese)
[19]	AUTAR K K. Mechanics of composite materials[M]. 2nd edition. Oxfordshire,UK: Taylor & Francis, 2006:87-90.
[20]	YANG Z Z, XU G M, YANG J Y, et al. Effect of different loads on the shoulder in abduction postures:a finite element analysis[J]. Scientific Reports, 2023, 13(1):9490. doi: 10.1038/s41598-023-36049-9
[21]	黄旭, 张世义, 李军. 图像分割技术研究综述[J]. 装备机械, 2021, 176(2):6-9.
	HUANG X, ZHANG S Y, LI J. Study overview of image segmentation technology[J]. The Magazine on Equipment Machinery, 2021, 176(2):6-9. (in Chinese)
[22]	刘燕萍, 郭佩瑶, 吴莹. 面向织物疵点检测的深度学习技术应用研究进展[J]. 纺织学报, 2024, 45(12):234-242. doi: 10.13475/j.fzxb.20240102302
	LIU Y P, GUO P Y, WU Y. Research progress in deep learning technology for fabric defect detection[J]. Journal of Textile Research, 2024, 45 (12):234-242. (in Chinese) doi: 10.13475/j.fzxb.20240102302
[23]	林国宇. 基于聚类的图像分割研究综述[J]. 电脑知识与技术, 2022, 18(26):17-18,24.
	LIN G Y. A review of image segmentation based on clustering[J]. Computer Knowledge and Technology, 2022, 18(26):17-18,24. (in Chinese)
[24]	胡群, 张宁, 潘如如. 应用均值漂移的印花面料交互式换色方法[J]. 纺织学报, 2021, 42 (11):97-102. doi: 10.13475/j.fzxb.20201103806
	HU Q, ZHANG N, PAN R R. Interactive color replacement method for printed fabrics based on Mean-shift[J]. Journal of Textile Research, 2021, 42 (11):97-102. (in Chinese) doi: 10.13475/j.fzxb.20201103806
[25]	TIAN K, LI J H, ZENG J F, et al. Segmentation of tomato leaf images based on adaptive clustering number of k-means algorithm[J]. Computers and Electronics in Agriculture, 2019, 165:104962. doi: 10.1016/j.compag.2019.104962 URL
[26]	HARTIGAN J A, WONG M A. Algorithm AS 136:a k-means clustering algorithm[J]. Journal of the Royal Statistical Society:Series C (Applied Statistics), 1979, 28(1):100-108.
[27]	何洁, 徐平华, 袁子舜, 等. 基于机器学习的UD布弹道冲击有限元结果分析[J]. 轻工机械, 2024, 42(4):7-15.
	HE J, XU P H, YUAN Z S, et al. Finite element results analysis of UD fabric ballistic impact based on machine learning[J]. Light Industry Machinery, 2024, 42(4):7-15. (in Chinese)
[28]	KURUMALLA S, RAO P S. K-nearest neighbor based DBSCAN clustering algorithm for image segmentation[J]. Journal of Theoretical & Applied Information Technology, 2016, 92(2):395-402.
[29]	TRAN T N, NGUYEN T T, WILLEMSZ T A, et al. A density-based segmentation for 3D images,an application for X-ray micro-tomography[J]. Analytica Chimica Acta, 2012, 725:14-21. doi: 10.1016/j.aca.2012.03.008 URL