Engine block type and position detection robotics application
A Mexican automotive supplier uses a robot as part of a fully-automatic system to transport engine blocks after the casting process into the processing station for breaking off and cutting of risers and gates. Six engine blocks at a time are pre-aligned on a 2.50 x 1.60 m transport pallet and conveyed through the cooling zone to the cutting station. To grip the parts, the robot requires the position of the individual engine blocks with an accuracy of +/- 2 mm. To prevent stress cracks due to uneven cooling, there is a chill casting around 40 cm long in the engine blocks. After removal of the engine blocks, the chill castings remain on the transport pallets and are then also removed using a robot at the next station. The position variance of the chill castings on the pallets is +/- 50 mm. They can be rotated by up to +/- 15°. An accuracy of +/- 2mm is also required for this robot positioning.
A total of 12 cameras and the CV-600 PC machine vision system from OCTUM carry out 2D position detection for the engine blocks and chill castings. Six cameras each are installed above the transport pallet in the engine block removal station and in the chill casting removal station respectively. One camera has to be assigned to each engine block or chill casting due to the high level of accuracy demanded relative to the 2.5 x 1.6 m size of the pallet. The algorithms used work exclusively with the part contour, as the reflective surfaces of the parts have a different pattern with each casting. The relevant taught-in contour is saved as vector data and searched for in the production cycle. The high-speed position determination accurate at sub-pixel level is extremely robust even with significant position variance and significant variations in lighting conditions. The system controller triggers image acquisition on the 12 cameras via the Profibus interface. These images are used to determine the position of the part and transform this to the co-ordinate system of the relevant robot. The position data generated is transferred directly to the robot via a serial interface and also to the PLC via Profibus. The directly transferred position data is used to control the robot, while the data transferred to the PLC is only used for release purposes. With the total amount of data to be processed by the PLC, this enables time savings to be achieved.
The difficult conditions with varying cast and metal surfaces, as well as the fluctuating brightness in the production plant, represent a considerable challenge for the machine vision system. Only contour-based part detection can reliably achieve the necessary level of automation throughout the plant. In addition, the powerfull software enables all 12 cameras to be evaluated with a single system.