Real-time defect detection of laser additive manufacturing based on support vector machine

Abstract

Laser additive manufacturing is an advanced digital manufacturing technology used to build or repair metal parts layer by layer. However, monitoring and in-process defect diagnosis lag behind advances in other key technologies, which makes product quality control a challenging problem. In this paper, a novel real-time monitoring system is proposed to automatically detect defects using principal component analysis and support vector machine. A camera is used in the image acquisition system to capture molten pool image. Ten molten pool features were extracted and principal component analysis was used to reduce the dimensions of the feature set. Support vector machine is used to build a classifier to detect defects in the deposited layer. The experimental results show that the SVM method can achieve high defect detection rate when identifying both slag and bulge defects. The support vector machine has a more satisfactory performance than the RBF neural network method. It is proved that the support vector machine method can be used more accurately and more universally in the in-situ monitoring system of laser additive manufacturing for defect diagnosis.