Ultrasonic indications in the roots of complete penetration welds (W1's) were the most common of the unexpected conditions identified after the 1994 Northridge earthquake in welded steel moment frame (WSMF) buildings. Widely treated by the engineering profession as earthquake-caused damage requiring repair, and characterized as such by the SAC, W1's have not previously been subjected to systematic study either to factually establish their ostensible relationship to earthquake ground motions or to debunk it. In this paper, the reliability of ultrasonic testing (UT) of full-penetration welded joints with backing in WSMF buildings is investigated. A study involving verification of the findings of UT via destructive investigation, a study of the influence of operator variability, and studies delineating some of the limitations of ultrasonics are described. In the first part of this two-part paper, several studies that were generally directed toward accurately characterizing W1's, identifying their origins, clarifying the extent and distribution of earthquake damage to WSMF buildings caused by the Northridge event, and developing recommendations for cost-efficient post-earthquake damage surveying of WSMF buildings were presented. Those studies found that W1's are not related to earthquake ground motion, and that the extent of Northridge earthquake damage to WSMF buildings is substantially less than previously published accounts suggest. In the present paper, it is shown that, as normally implemented by testing lab personnel, UT is an unreliable inspection technique for complete penetration T-joints with backing, which provides a plausible explanation for the widespread occurrence of W1's in existing buildings.
Read full abstract