Akademik Veri Yönetim Sistemi
Mining, Metallurgy and Exploration, cilt.42, sa.5, ss.2943-2962, 2025 (SCI-Expanded)
In this study, the performance of abrasive waterjet (AWJ) multi-pass and forward angling the jet cutting was experimentally investigated with the workpieces consisting of various rock types (igneous, metamorphic, and sedimentary), whose cutting depths were considered as the cutting performance indicator. Additionally, machine learning algorithms (MLAs) such as Gaussian process regression (GPR), artificial neural network (ANN), and support vector machine (SVM) were leveraged in modelling the cutting depths. Once the MLAs-based models were developed, moreover, the robustness of the models was comprehensively assessed through the metrics such as determination coefficient (R2), root mean square error (RMSE), and mean absolute percentage error (MAPE). In comparison to the single cutting, it was noted that rock cutting depths were significantly improved by the multiple pass cutting from at least 5.68 to 30.50% depending on the rock type. In addition, it was discovered that the forward angling the jet cutting improved the cutting depths of the limited rocks, compared with the cutting depths recorded at 90°. Furthermore, the GPR model showed the best performance in modelling the cutting depths with R2 of 0.98, RMSE of 2.665, and MAPE of 6.4% among the tested MLAs. Based on their R2, RMSE, and MAPE values, the ANN and SVM were ranked as second and third best models in modelling the cutting depths. Overall results demonstrated the superiority of the multi-pass cutting technique over the single-pass cutting and the effectiveness of the MLAs in modelling the cutting depths, achieved by the multi-pass and forward angling the jet cutting.