Machine learning-based predictive modeling for surface roughness in abrasive machining

The surface quality of any machined component or product greatly influences its attributes, including wear, corrosion resistance, fatigue strength, and more. Numerous safety-instrumented systems and other essential industrial systems necessitate components with a high-quality surface finish. Surface roughness (Ra) is one of the key measurements that indicate the finishing quality of machined parts. In this paper, we begin by detailing the specifics and Ra data set of an experimental study conducted on a commonly utilized abrasive machine, namely, a surface grinder. Next, machine learning algorithms are utilized for predictive modeling of Ra. The problem is framed using the Design of Experiments (DOE) to measure the Ra of EN 8 steel plates with two types of cutting fluids: a conventional synthetic cutting fluid and an eco-friendly cashew nut shell liquid (CNSL). A total of 19 experiments were performed, with five input variables established at two levels each. In the process of creating a predictive model, various algorithms, including Linear Regression, Decision Tree, Random Forest, Epsilon-Support Vector Regression (<mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"><mml:mi>ε</mml:mi></mml:math>-SVR), and K-Nearest Neighbors Regression, are utilized on the data obtained through experiments. Computed metrics, including Mean Square Error (MSE), Mean Absolute Percentage Error (MAPE), and Coefficient of Determination (<mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"><mml:msup><mml:mrow><mml:mi>R</mml:mi></mml:mrow><mml:mrow><mml:mn>2</mml:mn></mml:mrow></mml:msup></mml:math>), showed that <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"><mml:mi>ε</mml:mi></mml:math>-SVR outperformed all other methods. Therefore, further tuning of its hyperparameters was done. With this SVR model, predicted roughness values are displayed through a GUI.