【摘要】 粉末床激光选区熔化(LPBF)技术作为一种制造复杂轻量化复合材料零件的主流增材制造技术,被广泛地应用于航空航天、军工、医疗等领域,但翘曲严重影响薄壁工件打印的品质。开展薄壁翘曲的高速图像识别技术研究。通过有限元仿真模拟正常状态与翘曲情况下熔池形貌,使用高速相机,在线实时监测制造过程中熔池图像并提取熔池特征,通过特征参数对比与机器学习的方式进行基于熔池形貌特征的工件翘曲缺陷识别,实验结果表明该方法可以在翘曲初期充分识别出缺陷,准确率达94.8%,为建立反馈控制修正系统奠定了基础。更多还原

【Abstract】 Laser Powder Bed Fusion(LPBF) technology is a mainstream additive manufacturing technology for manufacturing complex lightweight composite parts. Such technology has been widely used in aerospace,military,medical and other fields. However,the warpage severely affects printing quality of thin-walled workpieces. This is a study on highspeed image recognition technology of thin-walled warpage. The finite element simulation was used to simulate morphological characteristics of the molten pool under normal and warping conditions. The high-speed camera was used to monitor images of the molten pool in the manufacturing process and extract characteristics of the molten pool in an online real-time manner.Through comparison of characteristic parameters and machine learning methods,the workpiece warpage defects were recognized based on morphological characteristics of the molten pool. Experimental results show that the defects have been recognized in the early stage of warpage,with an accuracy of 94. 8%,which lays a foundation for building a feedback control correction system.更多还原