article · Results in Materials
This research evaluates two modelling approaches, Response Surface Methodology and Artificial Neural Networks, for predicting and optimising performance during double tool turning of medium carbon steel without cutting fluids. Experiments carried out on an adapted lathe tested cutting speed, feed rate, and primary and secondary depths of cut to track surface roughness, material removal rate, and tool-tip temperature. The feed rate proved to be the primary factor governing surface finish, while cutting speed largely dictated material removal and thermal output. Both methods successfully mapped the non-linear links between operating settings and machining outputs. While the neural network achieved superior numerical precision with correlation values exceeding 0.97 and lower error rates, Response Surface Methodology offered clearer insights into how parameters interact. Together, the techniques provide a balanced route to enhancing productivity and surface quality in dry turning operations.
Dry machining reduces fluid costs and environmental hazards, but managing heat and component quality remains difficult. By demonstrating how machine learning and statistical models can accurately forecast surface quality, material removal, and temperature during double tool turning, this work helps manufacturing engineers set ideal operating parameters. This improves component finishing and tool durability without relying on costly trial-and-error physical testing.
This applied and tested experimental work is relevant to precision manufacturing workshops, automotive component producers, and lathe operators processing medium carbon steel. The modelling framework could be integrated into computer-aided manufacturing software or machine control interfaces to automate parameter selection for dry double tool turning. Because the experiments were completed on a modified conventional lathe, deploying the findings into industrial machine setups represents a mid-stage development effort focused on software integration and operational validation.
AI-generated from the published abstract. Always read the original work before citing.
This study presents a comparative evaluation of Response Surface Methodology (RSM) and Artificial Neural Network (ANN) approaches for predicting and optimizing machining responses during Double Tool Turning (DTT) of AISI 1043 medium carbon steel under dry machining conditions. The objective is to develop predictive models for surface roughness (Ra), material removal rate (MRR), and tool-tip temperature using experimental data obtained through a Box–Behnken Design (BBD). Experiments were conducted on a modified conventional lathe using four machining parameters: cutting speed, feed rate, primary depth of cut, and secondary depth of cut. RSM-based second-order polynomial models were developed and evaluated using Analysis of Variance (ANOVA), while ANN models were developed using a feed-forward multilayer perceptron with one hidden layer of ten neurons trained using the Levenberg–Marquardt algorithm. The model performance was assessed using R 2 , adjusted R 2 , mean absolute error (MAE), root mean square error (RMSE), and comparison between experimental and predicted responses. The experimental results indicated that surface roughness, tool-tip temperature, and material removal rate varied from 0.82–2.05 μm, 42–90.2°C, and 1.5–13.5 mm 3 /min, respectively. Feed rate was the dominant factor affecting surface roughness, while cutting speed significantly influenced MRR and thermal behavior. Both models successfully captured the nonlinear relationship between machining parameters and responses; however, ANN showed higher prediction accuracy with lower RMSE and MAE values and correlation coefficients above 0.97, whereas RSM provided superior interpretation of parameter effects and interactions. The results confirm that combining RSM and ANN offers an effective framework for machining response prediction and optimization, enabling improved surface quality, productivity, and thermal performance in Double Tool Turning operations.
This page summarises published work. The authoritative version sits with the publisher.
DOI: 10.1016/j.rinma.2026.101036
Is something wrong with this record? Report it or request removal.
Discussion
Have you built on this work, tried to replicate it, or seen it applied in practice? Share what you know. Verified researchers and MARATTO™ domain experts can open a discussion, and any member can reply. Contributions are reviewed before they appear.
No discussion yet. Open the first thread.
New to MARATTO™? Create a free account.
