Identification des paramètres modaux de systèmes mécaniques non stationnaires : application à un robot flexible de meulage

Abstract

This thesis is part of a research project launched by ETS and Hydro-Québec on the dynamics of a flexible mobile robot named SCOMPI which is used for repairing the hydraulic turbines on site. The main objective of this research project is to improve the quality of the robotic grinding process of SCOMPI, which is influenced mainly by the chatter and vibration during the process. To avoid such a problem, IREQ has conducted researches on the phenomenon of chatter and modeling of the process. Their research shows that the phenomenon of chatter depends on the modal parameters of the robot. So identifying the modal parameters of the robot is essential for their research. The study presented in this thesis focuses on identification of modal parameters of a flexible robot in operation based on operational modal analysis of non-stationary systems. Two identification methods are adopted in the thesis: AFS-VTAR model (Adaptable Functional Series Vector Time-dependent Autoregressive) and Short-Time SSI-DATA (Data-Driven Stochastic Subspace Identification). They actually represent the principal identification methods for time-varying system. One disadvantage of them is that the identified results contain both the expected modal parameters and computational ones. In fact, the number of methods to differentiate the modal parameters from the computational ones is still limited. The originality of this thesis is the development of two new indexes for AFS-VTAR and SSI-DATA for differentiating the real ones from the false ones. The effectiveness of the proposed indexes was validated through numerical simulations signals and experimental data from measurements on the flexible robot in operation. Recommendations for future research are proposed at the end.

Date	16 Nov 2016
Original language	French
Awarding Institution	École de technologie supérieure
Supervisor	Marc Thomas (Supervisor) & Zhaoheng Liu (Co-supervisor)