A robust prediction method of surface deflection using stoning simulation and curvature analysis

Recently, numerical predictions for surface deflection by finite element analysis have been carried out actively by various methods, such as curvature analysis, stoning simulation, and highlighting. Measures of surface deflection based on curvature analysis had been proposed by several researchers and have shown good correspondence with experimental results. The maximum variation of curvature difference between the panel and the tool was also proposed as a measure for surface deflection. However, the curvature is strongly dependent on the calculation method and the quality of data points. On the other hand, a measure based on stoning simulation has not been extensively studied yet due to the insufficient information from stoning simulation. Regardless of the lack of a reliable measure, stoning simulation has been widely accepted in the field owing to its robustness to noise, and easiness of use and understanding. In this paper, we propose a new robust prediction method that uses stoning simulation and curvature analysis sequentially. The method combines the merits of the two methods in an integrated approach. First, for detection and finding the approximate shape of surface deflection, stoning simulation was utilized. Next, for the more accurate quantification of surface deflection, curvature analysis was applied. We used an adaptive span method to calculate the curvature stably and accurately regardless of the noise. As a verification example, a shallow rectangular drawing with inner rectangular embossing was considered. Experimental and simulation results were compared, and they demonstrated that the proposed approach can be utilized for the reliable detection of surface deflection and the accurate quantification of surface deflection.