Autonomous Profile Tracking for Multiaxis Ultrasonic Measurement of Deformed Surface in Mirror Milling
Peng Guo, Limin Zhu, Zeming Wu, Wenze Zhang, Nuodi Huang, Yang Zhang
Abstract
High-precision remaining wall thickness is desirable in the mirror milling of large thin-walled parts. However, the workpiece deformation caused by sheet forming, clamping, and self-gravity makes it difficult to control the thickness if the original tool path generated from the designed CAD model is used directly. In order to solve this problem, the deformed workpiece surface is supposed to be reconstructed, and the original tool path should then be adjusted. In this article, based on a self-developed multiprobe ultrasonic measurement system, an autonomous profile tracking approach, mainly including adaptive sampling point prediction, boundary processing, and feature point extraction, is proposed to scan the deformed surface. The sampling point prediction algorithm calculates the position and orientation simultaneously for the next sampling point. In particular, a current-measurement-output-based orientation prediction algorithm is proposed to guarantee the autonomy of the profile tracking, and a reference-measurement-point-based orientation prediction algorithm is proposed to improve the prediction accuracy. The boundary processing algorithm controls the scanning direction and ensures that the sampling process is implemented inside the measurement boundary. The online feature point extraction algorithm reduces the redundant points for the following surface reconstruction. Practical experiments on freeform surfaces with/without complex boundaries verified the feasibility and effectiveness of the proposed method.