Smart Techniques of Microscopic Image Analysis and Real-Time Temperature Dispersal Measurement for Quality Weld Joints

The weld bead geometry examination by non-destructive testing plays a major role in macro to micro joining industries. Mostly, weld bead geometry revealed ample information associated with penetration depth and considered as major for weld quality joint. Instantaneous monitoring of weld bead quality acts as a major part of welding and internal control. Gas Tungsten Arc Welding (GTAW) is extensively operated in various industries. In general, weld bead superiority is firmly reliable and away from flaws by the optimum selection of welding process parameters and supplementary arrangements before welding is done. The impulsive aspects such as fluctuations of weld bead spot size and variability of welding penetration generate unsteadiness of welding excellence seeing complications of most arc welding processes. There are numerous limits such as voltage, current, gas flow rate, welding sound, and weld pool image meticulously connected to welding superiority. This data mainly contains voltage, current, gas flow rate, and geometry constraints. For finding the evidence and approximation of weld bead geometry, the author proposed promising image analysis techniques of Gray Level Co-occurrence Matrix (GLCM) for the extraction of statistical texture features. In the presented work, gas metal arc welding was carried out on 304L stainless steel plate with dimensions 300 mm × 150 mm × 1.5 mm. Experimental results confirmed the GLCM technique by extracting weld seam shape and also determining critical location information. At last, the GLCM revealed promising statistical texture features of weld bead geometry and found closer to experimental results. The experimental set-up is developed to measure the online bottom surface temperature of the workpiece during welding with the help of contact-type thermocouples located below the workpiece. For validation, a 2D and 3D heat conduction analytical model is used to predict penetration depth and the predicted depth of penetration is compared with experimentally obtained depth penetration.