Prediction of the geometric characteristics of the laser cladding of Inconel 718 on the Inconel 738 substrate via genetic algorithm and linear regression

Abstract

Assessing the geometric characteristics of a track is the first step in indicating the results of the laser cladding process. In this research, the effect of three essential process parameters, namely scanning speed, laser power, and powder feed rate, on the geometric characteristics of the Inconel 718 track on the Inconel 738 substrate have been studied. The goal of this research was to obtain the optimal process parameters for Inconel 718 laser cladding. To investigate the impact of process parameters on the geometric characteristics, five levels of laser power, three levels of scanning velocity, and three levels of powder feed rate have been considered, and a full factorial design of experiment has been performed. Consequently, 45 tracks have been cladded, and their geometric characteristics, including penetration depth, wetting angle, track height, track width, and compound characteristics such as dilution, has been investigated. After analyzing the results, the behavior of the geometric characteristics for the same process parameters in other brackets have been predicted via linear regression and genetic optimization via MATLAB software. PαVβFγ is participating in Y=aX+b as a compound variable. The regression process is feasible via accurate estimation of exponents α, β and γ. Therefore, the genetic algorithm was used to reduce the errors of linear regression. It is concluded that all of the process parameters have quantitative and qualitative effects on the geometry of the track. Finally, based on the modeling results, a process map was drawn to predict the impact of process parameters on the track geometry.