Statistical modelling and optimization of microhardness transition through depth of laser surface hardened AISI 1045 carbon steel

The mechanical properties of laser surface hardened material are closely related to the hardened layer. Respective study of geometrical dimension and microhardness would only result in insufficient understanding of the hardened layer. In this paper, a central composite design (CCD) is proposed, based on response surface methodology (RSM), to investigate the influences of processing parameters including laser power (LP), scanning speed (SS) and defocusing distance (DD) on the geometrical dimensions, microhardness and microhardness transition of laser surface hardened AISI 1045 carbon steel. Two-tailed Pearson correlation of selected responses was firstly undertaken to screen out unrelated ones. Second order response surface models were then developed and tested for three responses: hardened width (HW), microhardness of hardened area (MH) and hardness gradient (HG). Scanning electron microscopy(SEM) was conducted to evaluate the evolution of microstructure with three process parameters. The results indicate that three response models are capable of interpreting the hardened width, microhardness and hardness gradient with satisfactory accuracy. It is observed that the HW is only linearly determined by LP and SS, while the hardness properties (MH and HG) are both affected by all the investigated parameters significantly. The hardness of base metal can be improved from 200HV to 660HV after laser surface hardening. Quadratic correlation can be obtained when the HG is transformed inversely. Validation experiment was finally conducted for testing the generalisation of fitted regression models. As the maximum relative prediction errors for three responses are respectively 4.18%, 2.25% and −4.36%, good generalisation has been accomplished by the developed models.