Explore chapters and articles related to this topic
Classification and selection of sheet forming processes with machine learning
Published in International Journal of Computer Integrated Manufacturing, 2018
Elia Hamouche, Evripides G. Loukaides
In manufacturing, however, most of the recent work focuses on quantitative problems such as the prediction of springback mentioned above and cost estimation (Verlinden et al. 2008). Existing work on classifying part geometries tends to be rule based (Gupta and Gurumoorthy 2013; Kumar et al. 2017) and hence tedious and sensitive to past errors. Additional examples of ML in sheet metal forming applications are detailed in books by Kumar Nee (2017) and Dixit and Dixit (2008). The complex tasks that NNs can tackle make them an appealing choice for the purposes of classifying the geometries produced by metal forming. Similarly, NNs show promise in addressing the challenge of toolpath design in forming processes. Liu et al. (2015) evaluate a variety of NN architectures to optimise parameter control in the Incremental In-Plane Bending process. Opritescu and Volk (2015) train a NN to control a power-hammer device on the basis of desired output shape. In a follow-up paper (Hartmann, Opritescu, and Volk 2016), a more complete framework for such automation is provided.