In-situ Visual Monitoring Technology For Geometric Defects In The Laser Powder Bed Fusion Process

Laser Powder Bed Fusion(LPBF),as one of the most concerned technologies in additive manufacturing,is currently widely used in aerospace,medical and mold manufacturing fields.However,in the production process of the LPBF process,parts are prone to defects such as warpage,cracks,and dimensional errors.Even if the parts are manufactured using the best process parameters,such phenomena cannot be eliminated.And even if the defective part is repaired by post-processing,it cannot meet the requirements for use,and the cost of post-processing is usually more expensive.In order to solve the problem of lack of quality assurance for part in the laser powder bed fusion process,the process must be monitored.The geometric characteristics of the process,such as powder spreading condition,sintering condition of fusion area and deviation of parts processing dimensions,are particularly important for the quality of parts.To this end,this paper studies the 3D surface topography of the powder layer,the 3D surface topography measurement method of the forming layer,the measurement method of the 3D contour of the forming area and the geometric defect analysis method,and proposes an insitu visual inspection technology for geometric defects during the laser powder bed process.The research work and innovations of this article are as follows:The principle of 3D surface structure light measurement is studied.An enhanced phase measurement profilometry based on double constraints is proposed to replace the phase unwrapping,which can greatly reduce the number of images captured during the process and achieve fast and accurate 3D surface topography measurement.A parallax constraint based on layer thickness range is proposed to solve the problem of slower matching of corresponding points.A similarity constraint based on the normalized crosscorrelation(NCC)number method is proposed to solve the problem of mismatching of corresponding points.The experimental results show that our method can reduce the search range of corresponding points to dozens,and can effectively eliminate the interference of mismatched points,and achieve fast and high-precision measurement of the 3D surface topography of the powder layer and the forming layer.The measurement precision is no more 0.030 mm.Various contour extraction methods are studied.A mask-based level set method is proposed.A mask-based level set method is proposed,which greatly reduces the number of iterations of the level set method.The accuracy of this method can reach sub-pixel level,and the Dice index is 0.971.A fast contour extraction method based on phase guidance is proposed.The phase difference information is used to locate the forming area,which greatly reduces the influence of factors such as uneven image gray distribution on the contour extraction method,and realizes rapid and high-precision contour extraction.The contour extraction is completed within 1s,and the Dice index is 0.951.Finally,the constraint of the point-to-plane distance is studied to remove the erroneous corresponding point pairs of the contours in the left and right images,and the high-precision reconstruction of the threedimensional contour is completed.Comparing the 3D contour data calculated by the above method with the CAD model can characterize the dimensional processing deviation and manufacturing accuracy of the LPBF process equipment.Based on the above research,this paper developed a set of in-situ visual geometric defect detection system for laser powder bed process.A set of universal in-situ visual inspection device models are designed on the hardware,system calibration,the powder spreading layer 3D surface topography measurement,the forming layer 3D surface topography and 3D contour measurement,geometric defect analysis and data display are implemented on the software.Experiments in actual production show that powder spreading defects can be identified by the standard deviation(std)of the 3D surface topography of the powder spreading layer.When a powder spreading layer std> 0.100 mm,this layer has powder spreading defects.Warping defects can be identified by the number of points with Z> 0.3mm in the 3D data of the forming layer.The experimental results prove that the measurement method included in the system is effective,and the measurement accuracy and stability of the system are high,which can meet the production needs.