Optimization methodology of alternating current P-GMAW process by voltage-current signal analysis. Issue 1 (September 2016)

• Record Type: Journal Article
• Authors: Carvalho, Leonardo; Paiva, Anderson; Leme, Rafael; Costa, Sebastião; Aguiar, Thairone
• Is Part Of: International journal of advanced manufacturing technology. Volume 86:Issue 1/4(2016)
• Page Start: 565
• Page End: 580
• Publication Date: 2016-09
• Subjects: Design of experiments -- P-GMAW -- Non-destructive testing -- Signal processing; Manufacturing processes -- Periodicals; Production engineering -- Periodicals; 670.427
• DOI: 10.1007/s00170-015-8106-y
• Languages: English
• ISSNs: 0268-3768

Abstract

Abstract For maintenance and quality assurance during the execution of manufacturing processes, it is necessary to use careful experimental procedures to ensure customer satisfaction and high competitive standards in the market. This paper, which focuses on the class of non-destructive testing, proposes an optimization methodology for the alternating current pulsed gas metal arc welding (P-GMAW) process. The methodology is based on the numerical characterization of electrical signals of current and voltage by processing the obtained digital signals. The numerical characterization is performed by calculating the area under the curve of the electrical signals, which were then used to estimate the indicator difference relative percentage (DRP) of the signs of the experimental design for an experiment taken as a reference. Among the results, the study finds that the proposed method shows better results for signs composed of all its original wave frequencies at the expense of analysis performed on signals reconstructed from its three largest original wave frequencies. The results of the optimizations also showed more variability among themselves. In addition to the optimization analysis, the efficacy was evaluated in a T2 Hotelling control chart as a tool for control. Additionally, due to its application in multivariate processes, this study also analyzes the DRPs of electrical signals using statistical process monitoring.