2-D Matrix Phased-Array Examination of Advanced High-Strength Steel Resistance Spot Welds

Advanced high-strength steel (AHSS) spot welds have different failure modes determined by: stress state, embrittled areas and fracture properties, and defects such as pores, cracks, and lack of fusion.  Conventional destructive tests are insufficient for inspection of AHSS resistance spot welds.  Conventional nondestructive evaluation (NDE) techniques detect limited weld conditions, are very slow, and non-reliable.  New reliable NDE techniques are required for faster inspection and detection of detrimental flaw conditions, in particular “stuck weld” condition.  The objective of this work was to develop and validate 2-D matrix phased-array (PA) technology for reliable inspection of RSW.

Ultrasonic modeling and simulation was used to design high-frequency 2-D matrix PA probe having improved detection and sizing capabilities for inspection of AHSS resistance spot welds.  The prototype probes were later built and used as part of the experimental validation phase.  The difficulty of examining the welds was minimized to a great extent by use of advanced imaging software module and time, amplitude, and frequency domain algorithms.  Destructive cross sectioning, planar metallography, and peel tests were performed to validate the size of the welds and flaws.

A classification algorithm was developed that would identify stuck welds with a high degree of accuracy.  Experiments on thin gauge material (0.7 mm) showed that the software algorithm provided real-time identification of good versus stuck welds with only two incorrect classifications out of 200 samples.  Probability of detection (POD) of 98% was achieved for the thin gauge materials.  When operator input was added to the classification step, all 200 samples were correctly identified.  Similar results were obtained for thicker spot weld samples that were 2 mm to 2 mm stackups.  Examples of C-scan images from good and stuck welds are shown below.

In this work, the capabilities of the matrix probe and the proprietary algorithm for detection, imaging, characterization, and sizing were validated.  A good agreement between simulations predictions, UT PA experimental results, and destructive testing was achieved.

For more information, please contact Roger Spencer at rspencer@ewi.org or call 614.688.5216, or Menachem Kimchi at mkimchi@ewi.org or call 614.688.5153.