.png){kind=icon}
.png){kind=icon}
Return
Research on the prediction of wear distribution of ball-end mill based on milling GH4169 nickel-based superalloy
.png){kind=icon}
PRE
AI
Y
Y
L
L
DOI：10.1108/ILT-06-2025-0308.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
PurposeThis study aims to establish an accurate prediction model for nonuniform tool wear in GH4169 milling by integrating process optimization and intelligent learning techniques.Design/methodology/approachA two-stage approach was used: response surface methodology (RSM) optimized cutting parameters, and a Whale Optimization Algorithm-backpropagation (WOA-BP) neural network model was built using machining angle and time to predict localized tool wear.FindingsThe proposed RSM-WOA-BP model achieved high prediction accuracy, reducing root mean square error to 1.69 & micro;m and mean absolute percentage error to 1.14%, significantly outperforming conventional BP networks in robustness and generalization.Research limitations/implicationsBecause the model parameters are closely related to workpiece machinability, coating wear resistance and tool-workpiece contact geometry, significant changes in workpiece material, coating system, tool diameter, cutting-edge geometry or cooling/lubrication strategy may alter the wear mechanism and the angle-dependent load distribution, leading to systematic bias if the model is directly applied. In such cases, recalibration is required. The proposed workflow is transferable to other materials and tool/coating systems, provided that necessary recalibration and validation are conducted under the new conditions.Practical implicationsIn batch manufacturing, the machining parameters and tool type for a given operation are typically kept stable, so the calibration effort can be amortized over the production batch; the model can thus serve as a practical tool for process planning and wear monitoring.Originality/valueThis work integrates the strengths of RSM and WOA-BP to develop a high-accuracy model for predicting nonuniform tool wear in ball-end milling, ensuring both modeling precision and experimental efficiency. The model supports precise tool wear prediction in machining nickel-based superalloys with ball-end mills, enabling better control of tool life, cost reduction and improved reliability in complex aerospace and high-temperature applications.Peer reviewThe peer review history for this article is available at: https://publons.com/publon/10.1108/ILT-06-2025-0308/
Keywords:
Ball-end milling cutter
Wear distribution
Response surface methodology
Whale optimization algorithm
Journal
I
IF:
1.8
Papers:
103
Citations:
2.4K