Classification of Damages in Composite Material using Multi-support Vector Machine
参考中译:基于多支持向量机的复合材料损伤分类


          

刊名：Journal of The institution of engineers (India), Series C
作者：Rajiv B.（College of Engineering Pune）
Kalos Pritam（MIT Academy of Engineering）
Pantawane Prakash（College of Engineering Pune）
Chougule Vikas（MES College of Engineering）
Chavan Vishwanath（College of Engineering Pune）
刊号：710HA004-3
ISSN：2250-0545
出版年：2022
年卷期：2022, vol.103, no.4
页码：655-661
总页数：7
分类号：TH
关键词：Composite material；Damage；SVM
参考中译：复合材料；损伤；支持向量机
语种：eng
文摘：The mechanical properties of composite-reinforced natural fibres depend on parameters like fibre strength, length and chemical behaviour of fibre–matrix interfacial bond. The clarification of the research and development of composite-reinforced natural fibres improve the mechanical properties along with end applications. In this paper, multi-support vector machine (MSVM) is used to test the durability of composite material and the classification of damages in composite material. The images are collected across a composite material with 5–7?mm impingement. The images are filtered initially using an anisotropic filter. The classification is finally estimated out with MSVM classifier. Experimental validation is conducted over various composite materials, and the results are tested in real-time images. The validation shows that the proposed method attains the improved rate of accuracy in classifying the images than other existing state-of-the art models. Further, the durability of the material is tested in terms of material removal rate, wear resistance rate and material strength.
参考中译：复合材料增强天然纤维的力学性能取决于纤维强度、长度和纤维-基质界面粘结的化学行为等参数。阐明了复合增强天然纤维的研究和发展，提高了其力学性能和终端应用。本文利用多支持向量机(MSVM)对复合材料的耐久性进行检测，并对复合材料的损伤进行分类。图像是通过5-7毫米撞击的复合材料收集的。最初使用各向异性过滤器对图像进行滤波。最后利用最小二乘支持向量机分类器对分类结果进行估计。对各种复合材料进行了实验验证，并在实时图像中对结果进行了检验。验证结果表明，与现有的其他分类模型相比，该方法具有更高的分类正确率。此外，从材料去除率、耐磨性和材料强度方面测试了材料的耐久性。



国家科技图书文献中心  全球文献资源网  京ICP备05055788号-26  京公网安备11010202008970号  机械工业信息研究院 2018-2024