Estimation of hardened layer dimensions in laser surface hardening processes with variations of coating thickness

Abstract

In the laser surface hardening process, a laser absorbing coating should be applied on the workpiece in order to act as a laser beam energy absorber to efficiently absorb expensive laser power. Thus, a suitable choice of coating thickness as well as the coating material should greatly improve the quality of the hardened surface. However, very few researchers have systematically analyzed the effect of coating thicknesses on the quality of hardened surface. In this study, a series of laser surface hardening experiments were performed with various laser powers, laser travel velocities and coating thickness. In the hardening experiments, the SM45C steel was typically used and a 4 kW CO2 continuous wave laser resonator was used as the laser source equipment. The experimental results were analyzed in order to identify the general effects of coating thickness variations on the hardened layer dimensions; as a result, a range of optimum coating thickness were identified. According to the results, it is generally revealed that the optimum range becomes thinner as the hardening velocity increases. Based on the experimental data, a multiple regression modeling method and a neural network method were proposed to quantitatively describe the relationship between the hardening conditions and hardened layer dimensions. To design efficient estimation models, various architectures of regression models and neural networks were examined by estimation error analysis. The estimation results show that the neural network models are less sensitive to the architecture changes and can more accurately estimate the hardened layer dimensions than can the regression models. (C) 1998 Elsevier Science S.A.