A new model to distinguish welds performed by short-circuit GMAW based on FRESH algorithm and MLP ANN

The short-circuit gas metal arc welding has been continuously studied over the years, due to its important role in manufacturing processes. Concerning the process, many kinds of research are carried out aiming to understand the influence of the shielding gas in welds quality. In this context, this work treats the voltage and current welding signals as time series and applies a feature extraction based on scalable hypothesis tests, which is called FRESH algorithm, in order to obtain the signal features. After that, these features are applied in a multilayer perceptron artificial neural network, trained by the scaled conjugate gradient method, which classifies the welds according to the flow rate and type of shielding gas used in the process. The model presented excellent performance, which shows that the proposal is suitable to be used in welding quality monitoring.

[1]  Y. Benjamini,et al.  THE CONTROL OF THE FALSE DISCOVERY RATE IN MULTIPLE TESTING UNDER DEPENDENCY , 2001 .

[2]  Andreas W. Kempa-Liehr,et al.  Distributed and parallel time series feature extraction for industrial big data applications , 2016, ArXiv.

[3]  Chunli Yang,et al.  Effects of shielding gas composition on arc behaviors and weld formation in narrow gap tandem GMAW , 2017 .

[4]  P. Kah,et al.  Effects of shielding gas control: welded joint properties in GMAW process optimization , 2017 .

[5]  A. Koushki,et al.  Influence of shielding gas on the mechanical and metallurgical properties of DP-GMA-welded 5083-H321 aluminum alloy , 2016, International Journal of Minerals, Metallurgy, and Materials.

[6]  John Norrish,et al.  The controlled short circuit GMAW process: A tutorial , 2014 .

[7]  F. Massey.  The Kolmogorov-Smirnov Test for Goodness of Fit , 1951 .

[8]  Zoran Obradovic,et al.  Feature Selection Filters Based on the Permutation Test , 2004, ECML.

[9]  Martin Fodslette Møller,et al.  A scaled conjugate gradient algorithm for fast supervised learning , 1993, Neural Networks.

[10]  D Curran-Everett,et al.  Multiple comparisons: philosophies and illustrations. , 2000, American journal of physiology. Regulatory, integrative and comparative physiology.

[11]  Jorge Giron Cruz,et al.  A methodology for modeling and control of weld bead width in the GMAW process , 2015 .

[12]  Bibhuti Bhusan Biswal,et al.  Application of Artificial Intelligence Methods to Spot Welding of Commercial Aluminum Sheets (B.S. 1050) , 2014, SocProS.

[13]  Eamonn J. Keogh,et al.  The great time series classification bake off: a review and experimental evaluation of recent algorithmic advances , 2016, Data Mining and Knowledge Discovery.

[14]  Yannis Manolopoulos,et al.  Feature-based classification of time-series data , 2001 .

[15]  Nick S. Jones,et al.  Highly Comparative Feature-Based Time-Series Classification , 2014, IEEE Transactions on Knowledge and Data Engineering.

[16]  N. Mcpherson,et al.  Effect of shielding gas parameters on weld metal thermal properties in gas metal arc welding , 2015 .

[17]  S. Stehman.  Estimating the Kappa Coefficient and its Variance under Stratified Random Sampling , 1996 .

[18]  Ramez Elmasri,et al.  Implementation Options for Time-Series Data , 1997, Temporal Databases, Dagstuhl.

[19]  Jorge Giron Cruz,et al.  Modelling and control of weld height reinforcement in the GMAW process , 2018 .

[20]  Kristin Campbell.  What your distributior can offer you , 2007 .

[21]  Giovanni Mummolo,et al.  ANN modelling to optimize manufacturing processes: the case of laser welding , 2016 .

[22]  P. Welch.  The use of fast Fourier transform for the estimation of power spectra: A method based on time averaging over short, modified periodograms , 1967 .

[23]  Chunli Yang,et al.  Effects of shielding gas composition on arc properties and wire melting characteristics in narrow gap MAG welding , 2017 .

[24]  Jan Raethjen,et al.  Extracting model equations from experimental data , 2000 .

[25]  S. W. Campbell,et al.  Techno-economic evaluation of reducing shielding gas consumption in GMAW whilst maintaining weld quality , 2012 .

[26]  Norman McPherson,et al.  Artificial neural network prediction of weld geometry performed using GMAW with alternating shielding gases , 2012 .

[27]  Stan Szpakowicz,et al.  Beyond Accuracy, F-Score and ROC: A Family of Discriminant Measures for Performance Evaluation , 2006, Australian Conference on Artificial Intelligence.

[28]  Pierre Geurts,et al.  Pattern Extraction for Time Series Classification , 2001, PKDD.