Piston rings are critical components in the internal combustion engines of most vehicles, and their production is a complex and precise process. The quality control process is also essential to ensure that the rings meet the required standards and do not fail during use, and one of the quality control processes is visual inspection. This involves a visual examination of the piston ring, where trained operators examine these components to identify any cracks, scratches, or other surface defects. This study intended to develop an automatic inspection vision system in a production environment, that is able to accurately and efficiently detect defects, while improving the quality control process in the production of piston rings. This work involved the design and development of an automatic inspection vision system, which had to follow a series of specific requirements. These include an accurate manipulation of piston rings of different sizes, the use of high-resolution cameras and specific illumination to make it possible to acquire the surface’s defects (with diameters of approximately 0.3 mm), the development of specialized software to detect defects, and also the ability to maintain a cycle time of approximately 3 seconds per ring. The system was progressively tested: first the different modules were validated, and afterwards the final prototype was tested on a sample of piston rings. The overall accuracy, speed, and reliability of the system were evaluated. The automatic inspection vision system was found to be reliable to be used at a production level environment. It was possible to use it with piston rings of different diameters, heights, and materials. Also, it was possible to manipulate the rings without any deterioration. The preliminary results seem to indicate a good inspection accuracy, with the system being able to find the different types of defects, on all the considered surfaces. Due to the design of the overall system, which includes different stations that work in parallel, it was possible to achieve a reduced cycle time, with a compact machine. The system is also able to compute different statistical measures that, in the future, will help to obtain the production efficiency, and to find trends of piston rings production. The limitations of this work include the rings’ dimensions that can be used within the machine, since it was defined in the beginning to work only with rings with dimensions going from 70mm to 150 mm. Another limitation is that the system is not able to detect defects that are located in the inner surface of the piston rings. Since this is also a preliminary study, further validations are needed to ensure the correct behavior of the machine within a real production environment, with the use of a larger number of piston rings to fully evaluate the effectiveness of the system. This paper offers a novel and innovative approach to the inspection of piston rings, with the use of an automatic inspection vision system to improve the efficiency and accuracy of the quality control process. This system is able to work with piston rings of different dimensions and materials, providing a faster and more reliable alternative to manual visual inspection, allowing for the rapid detection and removal of defective products from the production line. In conclusion, the development of an automatic inspection vision system for detecting defects on piston rings represents a significant advancement in the field of quality control in the manufacturing industry. By combining machine vision techniques with machine learning algorithms, the system can quickly and accurately detect various types of defects, improving the efficiency and reliability of the inspection process. The use of this system could potentially reduce the need for manual inspection, freeing up valuable resources and reducing the risk of human error. Overall, this technology offers a promising solution for enhancing the quality control process in the production of piston rings, ultimately resulting in improved product quality and customer satisfaction.
Dr. Pedro Vieira, R&D Project Manager, Controlar - Innovating Industry