Explainable machine learning for abrasive water jet machining parameter optimization of Ti-6Al-2Sn-4Zr-2Mo alloy

Abrasive water jet machining (AWJM) of titanium alloys presents significant modelling challenges because the process responses depend on several interacting parameters in ways that simple regression cannot capture reliably. This paper builds a machine learning framework around XGBoost regression combined with SHapley Additive exPlanations (SHAP) to predict material removal rate (MRR), surface roughness (Ra), and kerf width in AWJM of Ti-6Al-2Sn-4Zr-2Mo alloy. An L18 Taguchi orthogonal array was used to generate 18 experimental runs across four parameters water pressure, abrasive flow rate (AFR), traverse speed (TS), and stand-off distance (SOD). Bayesian hyperparameter tuning gave the XGBoost model a cross-validated R 2 of 0.744 and 0.539 for MRR and Ra, respectively, outperforming the baseline models of linear regression and random forest. SHAP analysis revealed that the effects of water pressure and AFR are the most significant for MRR, while the effect of TS is dominant for Ra. All three responses were then plotted in a Pareto space where the trade-offs between the three responses were mapped and the non-dominated combinations of parameters were identified. The results demonstrate that interpretable ensemble learning is a viable approach for improving the parameter selection of advanced titanium alloy machining.

Authors

Institutions

Publication Details

Journal
Frontiers in Mechanical Engineering
Published
2026-11-17
DOI
https://doi.org/10.3389/fmech.2026.2007646
Primary Topic
Erosion and Abrasive Machining
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Explainable machine learning for abrasive water jet machining parameter optimization of Ti-6Al-2Sn-4Zr-2Mo alloy

Rajaram Prithivirajan, S. Rathnamala, S. Gejendhiran, A. Perumal et al.
4 links
Frontiers in Mechanical Engineering
Erosion and Abrasive Machining
article

Explainable machine learning for abrasive water jet machining parameter optimization of Ti-6Al-2Sn-4Zr-2Mo alloy

Rajaram Prithivirajan, S. Rathnamala, S. Gejendhiran, A. Perumal, T. Shalini
article en

Abstract

Abrasive water jet machining (AWJM) of titanium alloys presents significant modelling challenges because the process responses depend on several interacting parameters in ways that simple regression cannot capture reliably. This paper builds a machine learning framework around XGBoost regression combined with SHapley Additive exPlanations (SHAP) to predict material removal rate (MRR), surface roughness (Ra), and kerf width in AWJM of Ti-6Al-2Sn-4Zr-2Mo alloy. An L18 Taguchi orthogonal array was used to generate 18 experimental runs across four parameters water pressure, abrasive flow rate (AFR), traverse speed (TS), and stand-off distance (SOD). Bayesian hyperparameter tuning gave the XGBoost model a cross-validated R 2 of 0.744 and 0.539 for MRR and Ra, respectively, outperforming the baseline models of linear regression and random forest. SHAP analysis revealed that the effects of water pressure and AFR are the most significant for MRR, while the effect of TS is dominant for Ra. All three responses were then plotted in a Pareto space where the trade-offs between the three responses were mapped and the non-dominated combinations of parameters were identified. The results demonstrate that interpretable ensemble learning is a viable approach for improving the parameter selection of advanced titanium alloy machining.

Frontiers in Mechanical EngineeringVol. 12
SRM Institute of Science and Technology (IN), Indian Institute of Technology Kharagpur (IN)
Erosion and Abrasive Machining
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon