Development Of The Cladding Powder For Laser Direct Writing Signs And Predicting The Quality Of Clad

Laser cladding technology is a new type of material processing and surface modification technology, and laser direct writing system is based on signs of laser cladding, and with a welding robot to achieve a new sign written signs system. It is different with the traditional production process, which requires a more beautiful appearance, more corrosion resistance, better connectivity with substrate, so in order to better meet the sign requirements, developing special cladding powder for weathering steel Q450NQR1 is essential.In this paper, through explorations of uniform design, Special powder formulations for the system of laser direct writing was developed , the formulations contains the elements Ni, Cr, Si, B, Fe, etc., and its content is: Ni:20 ~22%; Cr:14 ~15.5%; Si:1 ~1.5%; B: 0.5~0.6%, Fe as the remainder. Using special powder without shielding gas during the laser cladding, the process is easier; and the cladding layer obtained has beautiful appearance, good atmospheric corrosion resistance and meets the requirements of the safe use of freight girder.Influence law of the laser cladding by parameters and the best process parameters were obtained by orthogonal experiments, and a comparison was taken between the cladding layer under the optimal parameters and Robot welding layer, which highlights the advantage of the laser cladding technology in the sign productionIn order to better achieve the promotion and application of laser cladding technique, multiple linear regression analysis was used to predict the quality of cladding layer. Regression equation indicated that the width and the height of the cladding, has a good linear relationship with the laser cladding parameters of the laser power, spot diameter, thickness of the powder and scanning speed, and the forecast accuracy of less than 7%, and sign in the regression equation has a clear professional meaning, so it has good predictive value.BP neural network based on the genetic algorithm has the higher prediction accuracy, which reaching 2.4%, but mapping is not clear, and the large amount of calculation, so it is only applicable to high precision applications.