Force control experiments for industrial applications: a test case using an industrial deburring example

摘要

Purpose - The main objective of this paper is to report the development of an indirect force control strategy designed to operate with industrial robotic deburring applications. More specifically, the system reported here is developed to debur high-quality knives that incorporate innovative design from well-known authors (fashion designers). Therefore, these products are very difficult to manufacture and have quality requirements incompatible with human-based deburring, since humans introduce too many unacceptable deviations as a consequence of their incapacity to maintain concentration for long periods of time. Design/methodology/approach - Since a good model of the environment is difficult to obtain, namely on industrial applications, a simple strategy was designed to obtain the relevant parameters leading to an acceptable performance. Consequently, the system implements an indirect force control strategy as a way to use actual robot controllers, explore the computing power of external personal computers, and the advanced features of modern force-torque sensors. The proposed strategy is presented in some detail and further discussed using a few test-case experiments. Findings - Experiments show a usable setup for contour following which is very useful to obtain the work-piece profile. A good selection of the path step seems to be, as expected, one of the most important variables to achieve good results: the smaller the increment over the trajectory the more regular is the resulting force profile. Low speeds also seem to lead to better results. The strategy implemented to maintain contact with the object and keep contact force at a certain level seems to result over surfaces with a smooth and large radius continuity, although there are significant force variations on impact with objects (which is not important since impacts can be planned), especially at the higher speeds, and even more significant near object edges. The desired contact force is also a parameter that should be tested. In the presented experiments, a contact force of 10 N was selected and oscillations of 1 N were observed around this value. In an industrial environment, more exposed to noise and vibrations, a higher contact force may be required. On the other hand, the increase of the contact force also increases the flexion of the sensing tool what brings more uncertainty to the calculated contact point. Large force oscillations imply more uncertainty of the obtained work-piece contour. Like in any industrial process selected parameters are the ones that show acceptable results at higher execution speeds. Practical implications - The objective of the presented setup is to find the better compromise for a particular industrial application, achieving acceptable operational cycle times. Originality/value - The obtained results are encouraging and the ability to perform contour recognition under a specified contact force can be very useful with the automatic deburring system being developed. In fact, this feature enables the system to acquire the exact contour of the working piece in the exact same conditions that will be used for the subsequent deburring task. This will contribute to minimize error and increase the process speed.