Predictive Model for Temperature-Induced Deformation of Robot Mechanical Systems.

摘要

The positioning accuracy and repeatability of a robot are critical for many industrial applications. Drift in repeatability can occur with changes in environmental and internal conditions, such as those seen with temperature-induced deformation. Thermal instability causes dimensional deformation, and a warm-up cycle is typically required to bring the robot to a thermally stable working condition. The elimination of warm-up cycles will ultimately enhance the positioning accuracy of the robots, their productivity, and reduce unnecessary energy consumption. The main objective of this research was to develop a robot controller algorithm that would provide, a priori, compensation for temperature-induced deformation associated with warm-up in robot mechanical systems. The research started at the fundamental stage of gaining insight into the thermal behaviour and corresponding temperature-induced deformation of simplified, i.e., one-dimensional, robot mechanical systems consisting of slender links and heat sources. The systems were studied using concomitant experimental, numerical and analytical models to provide cross-checking of the results. For the experimental model, the deformation was measured by tracking the drift of a laser diode spot across a charge-coupled device (CCD) camera chip. A non-contact measurement system consisting of an infrared camera, a CCD camera and a laser diode was developed to provide high accuracy measurement for the deformation. The numerical model was generated with a coupled thermal-mechanical finite element analysis incorporating thermal effects due to conduction and convection. The models were tested with the analytical model that was further extended using a finite difference technique. Once the three models showed excellent agreement, it was possible to develop a controller algorithm. Deformations predicted by the finite difference model were used as input for a validation experiment of the compensation algorithm. Results of the validation experiment confirmed that temperature-induced deformation of the simplified robot mechanical system can be accurately compensated for using a simple algorithm implemented in a robot controller.