A Feasible Method for Evaluating Energy Consumption of Industrial Robots

Establishing energy consumption models is important to achieve green manufacturing for robot automatic lines. To deal with the practical issue that the joint torque are difficult to be acquired in industrial robots, this paper aims to build dynamics model of robots from the motor torque to the robot motion. The Newton-Euler method is used to construct the model in linear-in-parameter form. In addition, as the conventional excitation signals, such as sine sweep and pseudo random binary signal, are not applicable for robots with closed control architectures, this paper proposes a method to use robot built-in point-to-point motion trajectory for the system identification. In this way, the parameters in the dynamics model are identified by linear least square. Eventually, energy consumption model can be formed. Simulation and experiments are conducted on a KUKA KR60-3 robot to verify the effectiveness of the proposed method. By experiment, the accuracy of prediction of energy consumption is as high as 90.54%.