Modelling of cutting forces and surface roughness evolutions during straight turning of Stellite 6 material based on response surface methodology, artificial neural networks and support vector machine approaches

Abstract

The present research work proposes an experimentalinvestigationhelpingto comprehend fundamental impacts of operating conditions during the cutting of cobalt alloys (Stellite6). Thus, an experimental design was adoptedto allowto build predicted mathematical models for the outputs, which are the average peak-to-valley profile roughness (Rz) and the tangential cutting force (Ft). Artificial neural network (ANN), support vector machine (SVM) and response surface methodology (RSM) were exploited to model the pre-cited outputs according to operation parameters. As a result, it hasbeenhighlighted that both feed rate and cutting depth, considerably, affect tangential cutting force evolution. Moreover, results show that both the insert feed rate and nose radius, are higher. Thismeans the averagepeak-to-valley profile roughness is higher. In order to put out the effect of operating parameters on cutting outputs, Analysis of variance(ANOVA)method has been employed. This has allowed the detectionof significant cutting conditions affecting average peak-to-valley profile roughness and tangential cutting force. In fact, to highlight the performance of adopted mathematical approaches, a comparison between RSM, ANN, and SVM has beenalso established in this study.

Keywords

• stellite 6
• turning
• rsm
• anova
• ann
• svm