Novel data driven modeling for cutting force prediction in CNC turning by machine learning approach

Prediction of cutting forces in CNC turning operations is an important aspect of modern manufacturing because it directly effects on tool wear, dimensional accuracy, surface finish, power consumption, and machining stability. Therefore, study presents a data-driven ML approach for prediction and analysis of cutting forces during CNC turning operations. Experiments were carried out on CNC turning machine using a TNMG160408EMTT cutting tool insert and D3 tool steel work piece material using bio-based cutting fluid. Cutting speed, feed, and depth of cut like machining parameters were considered as input parameter, while cutting force components F x (main cutting force), F y (feed force), and F z (radial force) were treated as output responses. The real-time acquisition of machining force data during experimentation is done by using Kistler Dynamometer. A total of 27 experimental runouts were performed, generating nearly 151,000 data points, which were further processed to obtain 27,000 effective samples for machine learning implementation. To enhance dataset quality & model reliability, the gathered dataset is subjected to pretreatment data operations such as missing value handling, standardization, and SMOTE (Synthetic Minority Over-Sampling Technique) balance. Arranged dataset is divided into 80% training section & 20% testing section. ML methods Linear Regression & Random Forest, Decision Tree were used to forecast cutting force and classify tool wear regression models are estimated using RMSE (Root Mean Square Error) and R² (coefficient of determination), while classification performance is evaluated by using F1-score with accuracy. According to experimental data, the classification model achieved nearly 87% accuracy and F1-score, while the suggested regression model produced R 2 score about 0.86. The created novel model and system can reliably identify tool wear situations and estimate machining forces. In Industry 4.0 & Industry 5.0 settings, the suggested ML based architecture can facilitate intelligent machining, tool condition monitoring, process optimization, and smart manufacturing applications. The research outcomes of the present work will be beneficial for the industrial applications to avoid the breakdown maintenance of CNC machines. The study was limited to a few conventional machine learning algorithms and did not include advanced models such as XGBoost, ANN, or deep learning methods.

Authors

Publication Details

Journal
Discover Mechanical Engineering
Published
2026-11-19
DOI
https://doi.org/10.1007/s44245-026-00420-7
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

Novel data driven modeling for cutting force prediction in CNC turning by machine learning approach

Udaysinh S. Bhapkar, Jambukeshwar Pujari, Akash Banne, Sunil Jamadade et al.
Discover Mechanical Engineering
Advanced machining processes and optimization
article

Novel data driven modeling for cutting force prediction in CNC turning by machine learning approach

Udaysinh S. Bhapkar, Jambukeshwar Pujari, Akash Banne, Sunil Jamadade, Harshal Mali, Vardraj Mahadik, Roshan Sutar, Sharad Patil
article en

Abstract

Prediction of cutting forces in CNC turning operations is an important aspect of modern manufacturing because it directly effects on tool wear, dimensional accuracy, surface finish, power consumption, and machining stability. Therefore, study presents a data-driven ML approach for prediction and analysis of cutting forces during CNC turning operations. Experiments were carried out on CNC turning machine using a TNMG160408EMTT cutting tool insert and D3 tool steel work piece material using bio-based cutting fluid. Cutting speed, feed, and depth of cut like machining parameters were considered as input parameter, while cutting force components F x (main cutting force), F y (feed force), and F z (radial force) were treated as output responses. The real-time acquisition of machining force data during experimentation is done by using Kistler Dynamometer. A total of 27 experimental runouts were performed, generating nearly 151,000 data points, which were further processed to obtain 27,000 effective samples for machine learning implementation. To enhance dataset quality & model reliability, the gathered dataset is subjected to pretreatment data operations such as missing value handling, standardization, and SMOTE (Synthetic Minority Over-Sampling Technique) balance. Arranged dataset is divided into 80% training section & 20% testing section. ML methods Linear Regression & Random Forest, Decision Tree were used to forecast cutting force and classify tool wear regression models are estimated using RMSE (Root Mean Square Error) and R² (coefficient of determination), while classification performance is evaluated by using F1-score with accuracy. According to experimental data, the classification model achieved nearly 87% accuracy and F1-score, while the suggested regression model produced R 2 score about 0.86. The created novel model and system can reliably identify tool wear situations and estimate machining forces. In Industry 4.0 & Industry 5.0 settings, the suggested ML based architecture can facilitate intelligent machining, tool condition monitoring, process optimization, and smart manufacturing applications. The research outcomes of the present work will be beneficial for the industrial applications to avoid the breakdown maintenance of CNC machines. The study was limited to a few conventional machine learning algorithms and did not include advanced models such as XGBoost, ANN, or deep learning methods.

Discover Mechanical EngineeringVol. 5(1)
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