A multilayer shallow learning approach to variation prediction and variation source identification in multistage machining processes

摘要: Variation propagation modelling in multistage machining processes through use of analytical approaches has been widely investigated for the purposes dimension prediction and variation source identification. Yet complex features is non-trivial task to model mathematically. Moreover, application associated identification techniques using Skin Model Shapes unclear. This paper proposes a multilayer shallow neural network regression approach predict geometrical deviations parts given manufacturing errors. The trained on simulated data, generated from simulation point cloud part. Further, data machined feature, can be identified by optimally matching deviation patterns actual surface with that surface. To demonstrate method, two-stage process virtual part planar, cylindrical torus was considered. geometric characteristics sources could predicted at an error 1% 4.25%, respectively. work extends analysis manufacturing.