Barriers for industrial implementation of in-process monitoring of weld penetration for quality control

Abstract

The research conducted sheds a light on the question why robust in-process monitoring and adaptive control are not fully implemented in the welding industry. In the research project FaRoMonitA, the possibilities to monitor the weld quality during welding have been investigated. Research conducted in this area has merely focused on technical issues investigated in a laboratory environment. To advance the research front and release some barriers related to industrial acceptance, the studies conducted in this paper have been both quantitative and qualitative in form of experiments combined with an interview study. The quality property weld penetration depth was chosen for in-process monitoring to evaluate the industrial relevance and applicability. A guaranteed weld penetration depth is critical for companies producing parts influenced by fatigue. The parts studied were fillet welds produced by gas metal arc welding. The experiments show that there is a relationship between final penetration depth and monitored arc voltage signals and images captured by CMOS vision and infrared cameras during welding. There are still technical issues to be solved to reach a robust solution. The interview study indicates that soft issues, like competence and financial aspects, are just as critical.

Monitoring of root gap change based on electrical signals of flux-cored arc welding using random convolution kernel transform

“…However, their usage is limited to predetermined welding conditions. Under unintentionally changing conditions, such as thermal distortion or machining defects of the workpiece, nonuniform weld quality and weld defects are easily found, which are detrimental to the fatigue life and strength of the weld joints [6][7][8][9]. Therefore, the workpiece must be monitored to ensure the quality when using the automated welding machine [10,11].…”

Quality Prediction and Control of Assembly and Welding Process for Ship Group Product Based on Digital Twin

“…Many scholars have conducted extensive research on welding quality prediction and control. Öberg and Sikström used arc voltage measurement, CMOS vision, and infrared camera to online monitor the quality characteristics of weld penetration and evaluated the industrial applicability and applicability of the framework [ 1 ]. Zhang et al developed a servo welding experimental system, which can extract the electrode indentation from the servo encoder to reflect the change of welding quality.…”

Artificial Neural Network Controlled GMAW System: Penetration and Quality Assurance in a Multi-Pass Butt Weld Application

“…Hence, the more stringent weld quality requirements of optimized structures, resulting from the use of thinner materials, high-strength steels and reduction of weld material, are difficult to achieve [1][2][3]. Additionally, the quality and penetration inconsistency, caused by the inability to adapt, affects drastically to fatigue life and strength properties of the weld joint [4][5][6][7]. Traditionally, meeting such quality requirements have been approached by precise machining and fitting of the workpieces.…”