Study on performance of machine learning models in predicting surface roughness when turning of hardened AISI 4340 steel using CERMET cutting tool
|
Full Text |
Pdf
|
|
Author |
Armansyah Ginting and Sudar Sono Sarjana
|
|
e-ISSN |
1819-6608 |
|
On Pages
|
1105-1111
|
|
Volume No. |
19
|
|
Issue No. |
17
|
|
Issue Date |
December 10, 2024
|
|
DOI |
https://doi.org/10.59018/092440
|
|
Keywords |
ra parameter, linear regression, random forest, support vector regression, Taguchi.
|
Abstract
This study investigates the effectiveness of machine learning models in predicting surface roughness (Ra parameter) during the hard turning of AISI 4340 steel using CERMET cutting tools, an area with limited prior research. The experiments were conducted on a CNC lathe under dry conditions, with surface roughness measured using a profilometer. The Taguchi design of experiments was utilized to structure the data, incorporating factors such as cutting speed, feed rate, and depth of cut. Three machine learning algorithms-Linear Regression (LR), Random Forest (RF), and Support Vector Regression (SVR)-were employed to develop predictive models. The models' performance was evaluated using the root mean square error (RMSE) and R-squared (R²) metrics. The RF model demonstrated superior predictive accuracy with an RMSE of 0.0758 μm and an R² value of 0.9814 for the testing dataset. The performance of the LR model (RMSE = 0.0783 μm, R² = 0.9791) and the SVR model (RMSE = 0.0779 μm, R² = 0.9798) were close to that of the RF model. Overall, all the algorithms demonstrated good potential for model development in this study.
Back