Tool path compensation strategies for single point incremental sheet forming using multivariate adaptive regression splines

Single point incremental sheet forming is an emerging sheet metal prototyping process that can produce parts without requiring dedicated tooling per part geometry. One of the major issues with the process concerns the achievable accuracy of parts, which depends on the type of features present in the part and their interactions with one another. In this study, the authors propose a solution to improve the accuracy by using Multivariate Adaptive Regression Splines (MARS) as an error prediction tool to generate continuous error response surfaces for individual features and feature combinations. Two feature types, viz.: planar and ruled, and two feature interactions, viz.: combinations of planar features and combinations of ruled features are studied in detail, with parameters and algorithms to generate response surfaces presented. Validation studies on the generated response surfaces show average deviations of less than 0.3 mm. The predicted response surfaces are then used to generate compensated tool paths by systematically translating the individual vertices in a triangulated surface model of the part available in STL file format orthogonal to the surface of the CAD model, and using the translated model to generate the optimized tool paths. These tool paths bring down the accuracy for most test cases to less than 0.4 mm of average absolute deviations. By further combining the MARS compensated surfaces with a rib offset strategy, the accuracy of planar features is improved significantly with average absolute deviations of less than 0.25 mm.

► MARS is a robust error prediction tool for incremental forming.
► Predictions can be made for individual features as well as feature interactions.
► A systematic method of generating compensated tool paths is now available.
► The compensation technique can be applied quickly with a factor of +1.
► The proposed technique improves the accuracy of parts significantly.