Experimental Characterization and Regression Modeling of Arc Height and Surface Roughness in Shot Peening

초록

The process of shot peening was modeled experimentally in terms of Almen arc height and the roughness of the treated surface using a multivariate regression framework. Experiments were conducted with controlled variations in the feed rate, orientation and position of the workpiece, and the resulting arc heights and surface roughness values were used to develop the models. The arc height model demonstrated strong predictive capability, yielding an R2 of 0.906 (modeling dataset) and 0.615 (independent validation dataset). However, surface roughness exhibited more complex variability due to localized material responses and random interactions; although statistically significant, the roughness model demonstrated moderate predictive accuracy. These results highlight that arc height can be estimated reliably from global process descriptors, whereas roughness prediction requires additional micro-scale information. The proposed approach provides an interpretable foundation for process optimization and supports the integration of data-driven enhancements for future shot peening control systems.

키워드