Physics-based modelling of robot's gearbox including non-linear phenomena

Digital Twin is a industry 4.0 technology highly deployed in control, safety and maintenance . Hence, increasing the digital twins' accuracy creates added value for these concepts. The digital representation of a robot requires the simulation of its dynamic behaviour. The friction torque in robot gearboxes seems to be a high considerable parameter in simulating robot dynamics. This paper presents a framework for representing the friction torque in industrial robot gearboxes. Initially, a literature review of the friction modelling approaches is conducted. The literature analysis indicates the LuGre model as an efficient approach for the description of the non-linear behaviour of the gearbox. The proposed procedure consists of a) the digital representation of the robot gearbox including LuGre friction model, b) the identification of LuGre model parameters with a metaheuristic Genetic Algorithm c) the validation of the proposed methodology using a 6-axis industrial robot. The digital representation of the robot gearbox is executed in OpenModelica. The LuGre parameters' estimation is achieved by defining and solving an optimization problem in Matlab and the utilization of data gathered from robot controller. The validation is based on comparing the torque response of digital and real word model for a specific position signal.