An Indicative End-Milling Condition Decision Support System Using Data-Mining for Difficult-to-Cut Materials Based on Comparison with Irregular Pitch and Lead End-Mill and General Purpose End-Mill

Data-mining methods using hierarchical and non-hierarchical clustering are proposed that will help engineers determine appropriate end-milling conditions. We have constructed a system that uses clustering techniques and tool catalog data to support the determination of end-milling conditions for different types of difficult-to-cut materials such as austenitic stainless steel, Ni-base superalloy, and titanium alloy. Variable cluster analysis and the K-means method were used together to identify tool shape parameters that have a linear relationship with the end-milling conditions listed in the catalogs. The response surface method and significant tool shape parameters obtained by clustering were used to derive end-milling condition decision equations, which were used to determine the indicative end-milling conditions for each material. Comparison with the conditions recommended by toolmakers demonstrated that our proposed system can be used to determine the cutting speeds for various difficult-to-cut materials.