Laser-Cladding Process Parameter Optimization of Fecocrnimn High-Entropy Alloy Based on Grnn and Nsga-Ii

Abstract

We would like to resubmit the enclosed manuscript to be published on “Optics & Laser Technology”. Thank you very much for giving us an opportunity for considering our manuscript. The further information is in the following:The TitleLaser-cladding Process Parameter Optimization of FeCoCrNiMn High-entropy Alloy based on GRNN and NSGA-IIAbstract: To enhance the performance of FeCoCrNiMn cladding layer, a study was conducted to optimize key variables, including laser power, scanning speed, and powder feeding rate. In addition, optimization variables such as the width-to-height ratio, dilution rate, and heat affected zone depth of cladding layer, were considered. To explore the effects of these variables and indexes, a three-factor, three-level full-factorial experiment was designed. A nonlinear model of the cladding parameters and cladding layer performance was established using the generalized regression neural network algorithm (GRNN). The predicted values obtained from this model exhibited an error of less than 10% when compared to the experimental results. Additionally, a set of optimal pareto fronts was obtained using the non-dominated sorting genetic algorithm (NSGA-II). The optimal combination of cladding parameters consisted of a laser power of 2.33 kW, scanning speed of 2.90 m/min, and powder feeding rate of 12.46 g/min. When the optimal parameters were used to prepare the cladding layer, it exhibited a compact section without cracks or pores. The combined use of the GRNN neural network and NSGA-II genetic algorithm yielded a favorable optimization effect. By judiciously selecting laser process parameters, it was proven to improve defects such as poor bonding force, multiple cracks, and pores in the cladding layer. Therefore, the wear resistance of the cladding layer was enhanced.The authorsDongya Zhang*, Jiaoyu Wu, Zhuoshen  Liu, Zhaoyang Zhai, Yanchao Zhang, Zhiqiang Gao

Keywords: Laser cladding, FeCoCrNiMn, Multi-objective optimization, GRNN, NSGA-II