Configuration similarity-based weak parts identification of full workspace for robot under the influence of milling excitation

Abstract

• By combining modal and cutting characteristics, a robotic weak parts identification method is established. • Configuration similarity index is proposed to extend the robotic weak parts identification method to the whole workspace. • The visualization of the distribution of the robotic weak parts is valuable for engineering practice. • The feasibility of the proposed method is verified by several experiments under three types of posture. Vibration is a major obstacle to the application of robot in the field of milling processing, and the low-frequency vibration of the robotic structure is likely to occur and have a severe impact. However, the complexity of the robotic structural modes and the posture dependence of the dynamic characteristics bring serious trouble to the vibration suppression of the robotic structure, and the milling excitation will further change the distribution of the robotic structural vibration, which together lead to the difficulty for engineers to locate the vibration-prone weak parts and realize the targeted vibration suppression. Therefore, firstly, this paper analyzes the weak parts of each mode of the robot based on the multi-joint and full-DOF mode shape (MFMS), and proposes the modal weak parts weight coefficient (M−WPWC) for quantitative evaluation. Furthermore, based on Kalman Filter (KF) method, this paper considers the change of the contribution of each mode to the end vibration caused by milling excitation, and proposes the modal contribution index (MCI) for quantitative evaluation. Finally, the configuration similarity index (CSI) is proposed to realize the identification of M−WPWC and MCI for any posture in the target space, and combining them, the robotic weak parts weight coefficient (R-WPWC) is proposed to quantitatively evaluate the distribution of robotic weak parts. Based on the proposed method, one standard posture and two non-standard posture are selected for the identification of R-WPWC, and the distribution of the robotic weak parts for the three postures is visualized and expressed in the form of parallel coordinate plots. The results show that the ω z DOF of the robot’s 2 and 3 joints are the weakest parts among the 36 joint-DOF of the robot. Finally, joint vibration test experiments are carried out to further verify this conclusion. Based on the method proposed in this paper, the distribution of robotic weak parts under any posture and working conditions is accurately identified, which lays the foundation for the formulation of targeted vibration suppression strategies.