Reconhecimento de cordões de solda para a inspeção automatizada de vasos separadores

• Main Author: Terres, Vinícius de Vargas
• Publication Date: 2022
• Format: Master thesis
• Language: por
• Source: Repositório Institucional da UTFPR (da Universidade Tecnológica Federal do Paraná (RIUT))
• Download full: http://repositorio.utfpr.edu.br/jspui/handle/1/30301
• topic: Solda e soldagem - Inspeção; Robôs industriais; Testes não-destrutivos; Detectores; Robôs - Sistemas de controle; Algorítmos computacionais; Solder and soldering - Inspection; Robots, Industrial; Non-destructive testing; Detectors; Robots - Control systems; Computer algorithms; CNPQ::CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO; Engenharia Elétrica

The weld bead inspection in separator vessels is a practice for maintaining of the equipment’s operation. It is usually performed by ultrasound techniques. The automation of the inspection is an alternative for the reduction of the time and resources spent. One solution for automation is the development of robotic systems for checking the integrity of weld beads. The proper positioning of the ultrasonic sensors and their displacement along the inspected section is one of the challenges during the implementation of weld bead inspection robots. This problem is solved by the development of a method for the identification and navigation along the weld bead. This work presents a novel solution for weld bead recognition, aiming at weld bead tracking. An algorithm is proposed to provide reliable weld bead detection from the position of the weld bead center. A technique for weld bead tracking has been developed. It is based on a weld bead line tracking controller and a module for the compensation of the uncertainty caused by gravity. At the end of the navigated section, an algorithm searches for a new weld bead to continue the inspection process. The identification and tracking of the weld bead were implemented on the Climbing Robot for Advanced Inspection. The experiments were conducted on an apparatus with several weld bead sections, which allow the evaluation of the tracking in different orientations. The tests show that the robot identified the weld bead and followed it in all simulated conditions, with a maximum error of 2.86 millimeters.