Institut des Systèmes Intelligents
et de Robotique





Tremplin CARNOT Interfaces



Thesis Title : Identification and physical simulation of a Stäubli TX90 robot during high-speed milling

Keywords : Robot manipulator, Modeling, Calibration, Kinematic and Dynamic parameters , Elastic Deformation, Simulation, Milling.

Date an Location :

Lundi 14 Mai à 10 heures
UPMC – Campus Jussieu, 4 Place Jussieu - 75005 Paris
Tour 65, 3ème étage, couloir 65-66, salle 304

Jury :

Grigore Gogu  Professeur des universités, IFMA ,
Saïd Zeghloul  Professeur des universités, LMS  ,
Philippe Bidaud                Professeur des universités, UPMC,
Stéphane Régnier            Professeur des universités, UPMC,
Faïz  Ben Amar                Maître de conférences , UPMC,
Nicolas Jardin                  Ingénieur chercheur, EDF,
Frédéric Hasnaoui           Chef de groupe, EDF ,

Abstract :

The researches presented in this thesis were carried out at EDF R&D. They aim to replace most of the physical tests by numerical simulations in order to optimize robotic maintenance processes that are performed especially on nuclear facilities. Its goal is, generally, to obtain productivity gains by increasing the speed of conducting operations that implements processes and to control results quality.

The considered processes are methods of high-speed milling. They require high precision trajectory tracking (spatial position and velocity), despite uncertainties in the structural parameters of the system conducting the process and while taking into account its intrinsic dynamic behavior (saturation, friction) and dynamic disturbances (gravity, physical interactions).

The developed work deals, in an experimental setting, with the identification problems of inertial and kinematic parameters and with the structural behavior (deformation and friction) of a Stäubli TX90 robot on various loads. These issues are discussed here in a real case while taking into account the constraints of implementation in an industrial context.

All the identified parameters were integrated into a dynamic simulator, which reproduces the control laws implemented on the industrial system and the trajectory generator, used to analyze the dynamic behavior of the system during the implementation of a high-speed milling process.

On this basis, the correction of the effectors’ trajectories can be done to minimize the deformities caused by these different sources of error.