Research On On-line Defects Detection Of Metal Additive Manufacturing

Laser metal additive manufacturing technology enables direct near-net shaping of complex metal parts,with unique advantages of short manufacturing cycle,high material utilization,and high process flexibility,however,various metallurgical defects are easily generated in the laser metal forming process,which seriously affects the mechanical properties of the parts.This paper presents an online infrared scanning detection method for metallurgical defects in laser metal deposition additive manufacturing.the heat conduction mechanism of the metal forming surface defects and the temperature distribution at the surface defects were systematically studied and analyzed,and the defect detection method verification experiments were performed,filtering noise reduction,defect recognition,etc.Which are performed on the collected temperature fluctuation curve.The simulation of the thermal field on the laser metal forming defect surface was performed.The mathematical model of heat conduction on the forming defect surface was established,and the relationship between the defect size and the temperature field on the forming surface was revealed.Using ANSYS Workbench thermal simulation analysis of titanium alloy specimens,the relationship between surface temperature field distribution and defect size variation was obtained.According to the simulation data,the relational expressions of peak heights of crack width,depth and temperature were fitted respectively.Experiment of Infrared scanning detection laser metal forming defect surface.An experimental system for infrared scanning inspection of defects was built.The system was used to detect the artificial cracks,pores and actual cracks on the surface of laser-formed specimens such as TC4 titanium alloy and GH4169 superalloy.In the experimental results,compared with the non-defect area,there is a significant temperature characteristic peak at the defect,which verifies the feasibility of the infrared detection method for metallurgical defects.The temperature data analysis and processing algorithm was studied.A detrending algorithm is used to detrend the temperature fluctuation curve to ensure that the detection system more effectively identifies defects.The noise reduction effects of the three algorithms are compared:the linear moving average method,the five-points-three smoothing method,and the wavelet threshold filter method.The results show that the wavelet hard threshold filtering method is the best to reduce the noise.A defect online detection and identification system was developed to realize the output and positioning of the peak temperature at the defect,which laid the foundation for the on-line targeted elimination of defects.