Defects Recognition in Selective Laser Melting with Acoustic Signals by SVM Based on Feature Reduction

2018 3rd International Conference on Advanced Materials Research and Manufacturing Technologies

材料科学;机械制造

Ye, D.S.^1,2,3 ; Fuh, Y.H.J.^2 ; Zhang, Y.J.^2 ; Hong, G.S.^2 ; Zhu, K.P.^3

Department of Automation, University of Science and Technology of China, Hefei

230026, China^1

Department of Mechanical Engineering, National University of Singapore, Singapore

117575, Singapore^2

Institute of Advanced Manufacturing Technology, Chinese Academy of Science, Changzhou

213164, China^3

关键词: Defect diagnosis;    Defects recognition;    Dimension reduction;    Discriminant models;    Feature reduction;    Fisher discriminant analysis;    Selective laser melting;    Training and testing;   

Others  :  https://iopscience.iop.org/article/10.1088/1757-899X/436/1/012020/pdf
DOI  :  10.1088/1757-899X/436/1/012020

学科分类：材料科学（综合）

来源: IOP

【 摘 要 】

Defects among the selective laser melting(SLM) part hinder the development of the SLM process. This work provides an approach to conduct the monitoring and defect diagnosis by support vector machines (SVM) model using extracted features from acoustic signals. After training and testing with the linear SVM model, the result from the Fisher discriminant analysis (FDA) feature reduction performs optimal compared with those from the original features and the principal component analysis (PCA) feature reduction. The melted state monitoring and classification can be realized by simple discriminant model of SVM with extracted features after dimension reduction. The proposed method can be applied in the SLM process monitoring and defect diagnosis by acoustic signals with generalization.