Start Date

2016 12:00 AM

Description

The ultrasonic control of steel welds is crucial in the oil and gas transportation industry. The state-of-the-art method is zonal focusing which consists in focusing the ultrasonic beam at specific distances using a wedge in order to check the integrity of specific zones in the weld. Advanced methods such as the total focusing method (TFM) give potentially better images in terms of signal to noise ratio and flaw resolution [1]. Moreover, a complete view of the weld is possible using TFM, contrary to zonal focusing. In this paper, we present a fast implementation of the TFM imaging for weld inspection using wedges. In particular, we propose a GPU implementation that tremendously accelerates the process and makes nearly real-time applications possible. The configuration with two media necessitates an optimization iterative procedure to estimate the proper times of flight, which is implemented in the GPU as well. We also present several skip modes in order to visualize different flaw orientations and locations in the weld (cap, root, etc.). First, we give imaging results on an aluminum block containing artificial flaws such as porosity or lack of fusion. We make a comparison between TFM and conventional imaging, and show that TFM gives better results in terms of image quality (SNR, resolution). Finally, we show results of a real girth weld inspection and demonstrate that the use of TFM is definitely interesting in this context.

Language

en

File Format

application/pdf

Share

COinS
 
Jan 1st, 12:00 AM

Inspection of Steel Welds Using Total Focusing Method Imaging

The ultrasonic control of steel welds is crucial in the oil and gas transportation industry. The state-of-the-art method is zonal focusing which consists in focusing the ultrasonic beam at specific distances using a wedge in order to check the integrity of specific zones in the weld. Advanced methods such as the total focusing method (TFM) give potentially better images in terms of signal to noise ratio and flaw resolution [1]. Moreover, a complete view of the weld is possible using TFM, contrary to zonal focusing. In this paper, we present a fast implementation of the TFM imaging for weld inspection using wedges. In particular, we propose a GPU implementation that tremendously accelerates the process and makes nearly real-time applications possible. The configuration with two media necessitates an optimization iterative procedure to estimate the proper times of flight, which is implemented in the GPU as well. We also present several skip modes in order to visualize different flaw orientations and locations in the weld (cap, root, etc.). First, we give imaging results on an aluminum block containing artificial flaws such as porosity or lack of fusion. We make a comparison between TFM and conventional imaging, and show that TFM gives better results in terms of image quality (SNR, resolution). Finally, we show results of a real girth weld inspection and demonstrate that the use of TFM is definitely interesting in this context.