Open loop control system and tools for data acquisition and estimation of the weld bead depth in GMAW process

Guillermo Alavarez Bestard¹; Renato Coral Sampaio¹; Sadek C. Absi Alfaro¹

¹ University of Brasília

Baixar PDF doi:10.20906/CPS/CBA2018-0829

Resumo

Gas Metal Arc Welding process is extensively used in many industries. The geometry of weld bead is very important and commonly used for quality validation, but an on-line measurement is challenging due to the extreme environmental conditions on the welding arc. These conditions limit the process automation since the control loop cannot be closed with classical instrumentation techniques. The system developed in this work, collects arc welding variables, depth of the weld bead and thermographic information of the molten weld pool. It integrated non-contact sensors with image processing algorithms and allow the real-time control in open loop. A novel algorithm for obtaining a dynamic model to estimate the depth of the weld bead was developed, based on these techniques. The estimators use a multilayer perceptron neural network and are optimized for embedded devices and real-time processing on a Field-Programmable Gate Array device. The results show that the system and proposed method is useful in industrial and research environments.

Palavras-chave: data acquisition; image processing; infrared termography; weld bead depth