Clarification of processing mechanisms and optimization of multiple properties through statistical analysis Akira Ogawa, 3rd year doctoral student, Graduate School of Business Administration, Mejiro University Takenori Takahashi, Visiting Professor, Keio University
Data that represents the processed shape is called profile data. Profile data is generally expressed in two-dimensional or three-dimensional Cartesian or polar coordinates. If the evaluation target is three-dimensional, multiple profile data on the surface and cross section can be obtained from one object. By analyzing profile data obtained from appropriately planned machining experiments using linear or nonlinear regression, it is possible to clarify the causal relationship between the processed shape and the machining factors. Furthermore, by combining it with specific technical information, it is possible to elucidate the machining mechanism. Once the mechanism is clarified, multiple profile data can be considered as objective functions, optimized using mathematical programming under appropriate constraints, and the optimal conditions can be fed back to the machining process to optimize multiple machining characteristics.
In this presentation, we propose a methodology for optimizing the machining process by analyzing machining shape profile data based on a statistical model using nonlinear regression in JMP and obtaining the optimal solution using mathematical programming.
