Identification of Correlations Between 3D Surfaces Using Data Mining Techniques: Predicting Springback in Sheet Metal Forming

Abstract

A classification framework for identifying correlations between 3D surfaces in the context of sheet metal forming, especially Asymmetric Incremental Sheet Forming (AISF), is described. The objective is to predict "springback", the deformation that results as a consequence of the application of a sheet metal forming processes. Central to the framework there are two proposed mechanisms to represent the geometry of 3D surfaces that are compatible with the concept of classification. The first is founded on the concept of a Local Geometry Matrix (LGM) that concisely describes the geometry surrounding a location in a 3D surface. The second, is founded on the concept of a Local Distance Measure (LDM) derived from the observation that springback is greater at locations that are away from edges and corners. The representations have been built into a classification framework directed at the prediction of springback values. The proposed framework and representations have been evaluated using two surfaces, a small and a large flat-topped pyramid, and by considering a variety of classification mechanisms and parameter settings.

Minimal Vertex Unique Labelled Subgraph Mining

“…The reported experiments were all conducted using real data taken from an AISF sheet metal forming application described in the introduction to this paper, more specifically the fabrication of flat topped pyramid shapes made out of sheet steel. This shape was chosen as it is frequently used as a benchmark shape for conducting experiments in the context of AISF [9].…”

Research on the angle measurement system of bending workpieces based on multi-line structured light

“…Researchers have developed metal bending rebound prediction models for compensating the CNC bending parameters 8 – 11 There are many factors that affect the amount of rebound, including the size, shape, thickness, material, and structure of workpieces, as well as manufacturing parameters 12 , 13 . The bending rebound model cannot cover all of the factors affecting rebound at the same time.…”

Classification and selection of sheet forming processes with machine learning

“…Springback is a common problem in metal sheet forming processes and compromises the accuracy of finished parts. Springback occurs due to the natural elastic recovery of metal after the workpiece is released (Garcia-Romeu, Ciurana, and Ferrer 2009), and has not been reliably captured by Finite Element models (El-Salhi et al 2012). In a number of studies (El-Salhi et al 2012;Khan et al 2015;Kazan, Fırat, and Tiryaki 2009), various geometry representation methods and ML classifiers such as Support Vector Machines (SVMs) and NNs were used to predict springback.…”

“…Springback occurs due to the natural elastic recovery of metal after the workpiece is released (Garcia-Romeu, Ciurana, and Ferrer 2009), and has not been reliably captured by Finite Element models (El-Salhi et al 2012). In a number of studies (El-Salhi et al 2012;Khan et al 2015;Kazan, Fırat, and Tiryaki 2009), various geometry representation methods and ML classifiers such as Support Vector Machines (SVMs) and NNs were used to predict springback. When compared with empirical data, the ML classifiers outperformed the FEM models and simple generic classifiers.…”

“…Many of the geometry representation methods summarise the geometry information as a collection of the main geometric properties as op-posed to plain point clouds. In work by El-Salhi et al (2012), two different geometry representation methods were implemented and compared. The Local Geometry Matrix (LGM) is a global matrix made up of 3 × 3 local matrices with the point of interest at the center of the matrix.…”