Application of image processing to radiographic image for quantitative assessment of friction stir welding quality of aluminium 2024 alloy

Abstract

This work is an attempt to utilize X-ray radiographic image for evaluating friction stir welding process quantitatively rather than qualitatively. The X-ray image from radiographic test is quantified by applying image processing to estimate defect area in terms of number of pixels and compared with mechanical properties of weld joint. Taguchi method is used to design the experiment with input process parameters rotational speed, welding speed and tool offset value. The area of defect from X-ray image and tensile strength of weld joint possess inverse relation. According to Taguchi method, the process parameters that produce minimum area of defect and maximum tensile strength are found to be rotational speed of 1000 rpm, welding speed of 80 mm/min and tool offset at 0 mm from center line. The ANOVA result depicts that welding speed is the dominating parameter for both area of defect and tensile strength followed by rotational speed and tool offset.