An Interpretable Machining Performance Index for Tribology‐Informed Evaluation of CFRP Milling Using Explainable Machine Learning

Carbon fiber‐reinforced polymer (CFRP) machining involves a challenging trade‐off between productivity and machining‐induced damage. To provide a unified evaluation framework, this study proposes a Machining Performance Index (MPI) that combines normalized material removal rate (MRR) and penalized delamination into a bounded composite descriptor representing both productivity and damage performance. A dataset comprising 640 CFRP milling observations was analyzed under controlled experimental conditions. To ensure methodological rigor, machine‐learning models were trained exclusively using independent process parameters, namely, spindle speed, feed rate, and tool diameter, thereby avoiding target leakage. Normalization parameters were determined solely from the training subset and subsequently applied to unseen test data. Several regression algorithms were evaluated, with random forest providing the best balance between predictive accuracy and generalization capability. Model interpretability was investigated using SHapley Additive exPlanations (SHAP) and partial dependence plots (PDPs). The results identified tool diameter as the most influential predictor of MPI, whereas feed rate and spindle speed exhibited nonlinear and interaction‐dependent effects. Because MPI is derived from productivity and damage responses, the reported importance patterns should be interpreted as explanations of model behavior rather than direct evidence of physical mechanisms. Weight‐sensitivity and Pareto‐consistency analyses indicated stable MPI rankings across alternative preference scenarios and general alignment with balanced productivity–damage trade‐off regions. Ablation analysis showed that MPI represents a more complex learning objective than MRR alone. Overall, MPI provides a transparent framework for comparative assessment, data‐driven analysis, and tribology‐informed decision support in CFRP milling.

Authors

Institutions

Publication Details

Journal
Journal of Tribology
Published
2026-11-18
DOI
https://doi.org/10.1002/jtrb.70019
Primary Topic
Advanced machining processes and optimization
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

An Interpretable Machining Performance Index for Tribology‐Informed Evaluation of CFRP Milling Using Explainable Machine Learning

Zihni Alp Çevik, Gencay Sarıışık
Journal of Tribology
Advanced machining processes and optimization
article

An Interpretable Machining Performance Index for Tribology‐Informed Evaluation of CFRP Milling Using Explainable Machine Learning

Zihni Alp Çevik, Gencay Sarıışık
article en

Abstract

Carbon fiber‐reinforced polymer (CFRP) machining involves a challenging trade‐off between productivity and machining‐induced damage. To provide a unified evaluation framework, this study proposes a Machining Performance Index (MPI) that combines normalized material removal rate (MRR) and penalized delamination into a bounded composite descriptor representing both productivity and damage performance. A dataset comprising 640 CFRP milling observations was analyzed under controlled experimental conditions. To ensure methodological rigor, machine‐learning models were trained exclusively using independent process parameters, namely, spindle speed, feed rate, and tool diameter, thereby avoiding target leakage. Normalization parameters were determined solely from the training subset and subsequently applied to unseen test data. Several regression algorithms were evaluated, with random forest providing the best balance between predictive accuracy and generalization capability. Model interpretability was investigated using SHapley Additive exPlanations (SHAP) and partial dependence plots (PDPs). The results identified tool diameter as the most influential predictor of MPI, whereas feed rate and spindle speed exhibited nonlinear and interaction‐dependent effects. Because MPI is derived from productivity and damage responses, the reported importance patterns should be interpreted as explanations of model behavior rather than direct evidence of physical mechanisms. Weight‐sensitivity and Pareto‐consistency analyses indicated stable MPI rankings across alternative preference scenarios and general alignment with balanced productivity–damage trade‐off regions. Ablation analysis showed that MPI represents a more complex learning objective than MRR alone. Overall, MPI provides a transparent framework for comparative assessment, data‐driven analysis, and tribology‐informed decision support in CFRP milling.

Journal of Tribology
Usak University (TR), Harran University (TR)
Advanced machining processes and optimization
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