Prediction of cutting parameters and reduction of output parameters using machine learning in milling of Inconel 718 alloy

The present study focuses on the effects of cooling/lubrication conditions and cutting parameters on energy consumption (EC), carbon emissions (CE), surface roughness (Ra), cutting temperature (T), tool wear (Vb) and vibration (Vib) in sustainable milling of Inconel 718 alloy. Also, it was aimed to estimate the EC, CE, Ra, T and Vib values obtained in milling experiments using three different regression-based machine learning (ML) models. The performances of the models used in ML were compared using R-squared, MSE and MAPE performance criteria. In the experiments conducted by reducing the feed rate and cutting speed and in MQL machining conditions, it was observed that EC and CE values reached minimum values. In MQL machining conditions, it was observed that the lowest Ra values were achieved at high cutting speed and low feed rate. The lowest Vb was measured at low cutting speed and feed rate in air machining conditions. Increasing the cutting speed and decreasing the feed rate in MQL machining conditions had a positive effect on Vib. In the MQL machining condition, at 40 m/min cutting speed and 0.06 mm/rev feed rate, the lowest energy consumption and carbon emission were 0.76 kJ/s and 0.54796 kg-CO₂ respectively. The lowest surface roughness and vibration values were measured as 0.234 μm and 1.91 mm/s respectively, at 80 m/min cutting speed and 0.06 mm/rev feed rate in MQL machining condition. The lowest cutting temperature was measured as 31 °C at a cutting speed of 40 m/min and a feed rate of 0.06 mm/rev under air machining conditions. It was seen that the EC, CE, Ra, T and Vib values arising from the input parameters in the machining of Inconel 718 alloy could be successfully predicted using three different regression-based ML models.