CBN cutting tool's surface roughness and tool wear prediction using JOA-optimized CNN-LSTM

Area of Science:

• Materials Science and Engineering
• Manufacturing Technology
• Artificial Intelligence

Background:

• Nickel-based superalloys like Inconel 718 present significant machining difficulties due to poor thermal conductivity and work hardening.
• These challenges lead to rapid tool wear and compromised surface quality during machining operations.
• Existing predictive models often lack the accuracy and real-time capabilities required for complex materials.

Purpose of the Study:

• To develop and optimize a hybrid deep learning model for real-time prediction of surface roughness and flank wear.
• To address the machining challenges associated with Inconel 718 under Minimum Quantity Lubrication (MQL) conditions.
• To integrate the predictive model into MATLAB/Simulink for practical industrial deployment.

Main Methods:

• A hybrid Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM) deep learning model was developed.
• The model was optimized using the Jende Optimization Algorithm (JOA) and trained on data from 27 full-factorial machining trials.
• Data preprocessing involved normalization and outlier removal using IQR and Z-score methods.

Main Results:

• The JOA-optimized CNN-LSTM model achieved high prediction accuracy with R=0.9991, RMSE=0.0095, and MAPE=2.21%.
• The model demonstrated superior performance compared to conventional methods like SVM, ANN, and ANFIS.
• The model effectively captured complex nonlinear interactions between machining parameters and responses.

Conclusions:

• The developed hybrid deep learning model provides a robust and accurate solution for real-time monitoring of machining processes for hard-to-cut materials.
• Integration with MATLAB/Simulink enables real-time deployment, digital-twin compatibility, and scalability for smart manufacturing.
• This sensor-free approach offers a cost-effective and scientifically interpretable method for predictive maintenance and process optimization.