Challenge: Developing 3D machine vision system for robotic welding path adaptation to discrepancies between successive product geometries. Ensure 10-fold time saving as compared to manual path adjustment. Compensate for large geometric deviations within 50 × 40 mm operating window. Detect standard topologies of geometric and closed butt joints (slot >= 0.05 mm), all with one system. Upgrade robotic welding equipment for the production of stainless processing equipment of the highest quality. Enable simple use and quick adaptability to the new product geometries and joint topologies. Perform robust operation on stainless surfaces, matte or polished, in the presence of defects, scratches, dissimilar treatment of different joint sides surfaces (e.g. a combination of matte and polished side). Ensure industrial level of protection against electromagnetic interference when TIG welding. Implement functionality of joint finding and joint guidance.
Solution: With the development of machine vision system for adapting robotic trajectories to geometric deviances in the welding process, we were able to successfully meet all the challenges. The first step towards adaptive welding is a simple upgrade of robotic manipulator with a machine vision system, followed by learning the reference geometry by using the imaging system.
For a specific type of a product, only one learning, which is generally faster than the manual due to its contactless nature, is needed. When learning, the reference geometry is divided into any optimal number of welding paths and depending on the customer's experience. Each path is assigned a joint type. Points along the path are found using the imaging sensor on the robotic arm. The entire preparation process is done exclusively by using the robotic interface. With this, the system is ready for automatic serial welding.
National Instruments development platform, used in this process, enables fast and effective adaptation of the system features for functional and cost-optimized satisfaction of specific requirements.
Author: Robert Modic