Journal of Materials in Civil Engineering

Cementitious materials have been used in construction for several millennia and through several civilizations. Their use has been revitalized in the recent century, and they have become the major construction material for housing and infrastructure. The amalgamation of the science and engineering of these materials in recent decades has advanced the state of the art in their understanding and development. In a new century where sustainable construction is becoming an overriding consideration, the new scientific concepts can have direct impact on the way we use and formulate cementitious materials. On top of this, the scientific advances have provided the means for the potential use of the current portland cement as a platform to develop composite formulations of a high level of performance equivalent to and in some respects exceeding those of steel and ceramics. This may provide the basis for advanced and efficient construction techniques. The present article reviews the past, present, and future of cementitious materials and highlights the need for a comprehensive approach to maximize the advantages of the newly emerging cementitious materials and concretes.

Full-scale welded steel moment-resisting frame (WSMF) connections have been tested to failure at the University of Michigan and at Lehigh University. The tests were conducted at ambient temperature and under a simulated seismic loading protocol developed by the SAC Joint Venture, to determine if newly detailed unreinforced fully restrained post-Northridge connections can behave satisfactorily in future earthquakes. This paper presents a failure analysis of 16 full-scale WSMF connections. The effects of weld-access hole geometry and the roughness of its flame-cut surface are analyzed. The analysis describes the initiation and propagation of fatigue cracks at intersections of the weld access hole with the beam flange and from the flame-cut surface of the weld access hole in the WSMF connections. The paper suggests changes to the weld access hole, and the surface roughness for the flame-cut surface of the weld access hole, which would improve the performance of unreinforced post-Northridge WSMF connections subjected to seismic loads.

T-joint specimens were tested in tension to investigate the ductility and strength requirements for the base metal; particularly with heavy section wide-flange columns as the transverse member loaded in the through-thickness direction. The column sections were obtained over a range of years from four steel mills (integrated and electric furnace) under four different ASTM specifications. Sulfur levels ranged from 0.003 to 0.043% and carbon from 0.05 to 0.15%. The through-thickness strength of the column flanges exceeded 690 MPa, well above any possible demand that could come from ordinary structural steel. Because there is no yielding of the column flanges in the through-thickness direction, the lower ductility in this direction should not be a concern, even if the sulfur levels are at the upper limits of the ASTM A992 specification.

Fracture of unreinforced welded moment frame connections subjected to simulated seismic loads was caused by the initiation of fatigue cracks and their propagation to critical size. The fatigue cracks initiated at the web-to-flange intersection at the weld access hole, the valleys on the flame-cut weld access hole surface, the weld toe, and weld imperfections. Final fracture occurred when the fatigue crack extended unstably either in the base metal or in the weld metal. Final fracture is determined by the size of a crack, the stresses and strains acting on the crack, and the fracture toughness of the material. This paper presents the methodology used to establish the necessary and sufficient fracture toughness requirement for weld metal used in seismic applications. The methodology was based on fracture mechanics principals and on empirical correlations. The proposed Charpy V-notch (CVN) toughness is 40 ft-lb at 70°F and 20 ft-lb at 0°F for components subjected to +50°F and higher. This CVN requirement should preclude weld metal toughness from being a contributing factor to the fracture of unreinforced moment frame connections. Further improvements in the fracture performance of the connections must be accomplished by changes in design, detailing, fabrication, and inspection.

Fracture mechanics can be used to determine allowable flaw sizes for welds in steel structures. Models are developed, both for pull-plate tests and for building connections. Building connections are more severe because of stress concentration in the center of the weld length and column bending. Ultrasonic acceptance criteria are described, based on size of flaw, for surface and embedded flaws.

This work addressed two issues pertaining to the ultrasonic inspection of steel-moment-frame buildings that may have suffered damage as a result of the 1994 Northridge earthquake ground motion. These were: (1) Are the AWS D1.1 Code procedures reliable enough to detect flaws in the structural weldments of interest? and (2) Are there alternative Code-supplemental procedures that can be applied manually for not only improved flaw detection, but also flaw location, classification, and sizing? Twelve mockup specimens containing a total of 17 intentional worst-case-scenario flaws in representative beam-to-column weld geometries were fabricated. The Code-based detection results were marred by four misses and more than ten false alarms. In contrast, use of Code-supplemental procedures assured that no intentional test flaws were missed and false alarms were kept to a minimum. Only the Code-supplemental advanced procedures were used to characterize (as to type, location, orientation, depth, and length) a detected test flaw. In 95% of the measurements, the XYZ flaw location, length, and depth estimated were within 5 mm (0.2 in.) of the intended values.

Following the Northridge earthquake, the quality and accuracy of the ultrasonic testing (UT) of structural steel connections, and the abilities of technicians to perform such testing, was drawn into question. In establishing new weld quality acceptance criteria and appropriate inspection guidelines, it was important to validate the ability, or lack thereof, of ultrasonic testing technicians to accurately characterize and locate weld flaws in a variety of welded joint types. Round-robin testing of mock-up specimens representative of a variety of common joints used in steel moment frames, fabricated with embedded discontinuities of known size, location, and orientation, were examined by a number of ultrasonic testing technicians. The inconsistencies noted in the test results indicate a need for more specific UT procedures when used for weld acceptance, and a need for validation of the skill level of the UT technicians in their ability to perform the inspections.

This paper describes static small component pull and bend tests carried out by the Dynamic Load Weld Task Group on 2-inch-thick specimens using weld overlays. The tests were successful in demonstrating that weld overlays have the capability of accommodating the full static ultimate demands of the parent metal.

This paper presents recommendations on the repair of cracked columns based upon the writers’ experiences, found from repair of cracked column flanges following the 1994 Northridge Earthquake. This paper also describes tests carried out using volumetric weld repairs on thick plates to evaluate acceptable procedures for the repair of column flanges with partial cracks damaged from earthquakes. Cracks were often found in the bottom flange of welded joints. In some cases, cracks propagated into column flanges, and in a few cases into their webs. Repairs of cracked welded joints are discussed at some length in other documents and thus are not part of the scope of this paper. One method for the repair of cracked columns involves replacement of the column flange with a new plate. Another method of repair involves the use of volumetric welds to replace the cracked portion of the column flange. This paper includes a description of the development of procedures and limited testing to assess the adequacy of the volumetric repair method of partial column cracks.

A high degree of local bending was observed in the beam flange of a beam-to-column connection modeled with 3D finite-element analyses. The regions of high stress concentration were shown to occur at the weld access hole and the intersection of the bottom beam flange and column flange. The use of a specially designed part termed a “stress diffuser” with a beam-to-column connection, in which the beam web and bottom beam flange are separated near the weld access hole, was found to eliminate the high stress concentration at the weld access hole and reduce the high degree of local bending in the beam flange by 49%. The maximum principal stress at 2.54 mm from the column face was reduced by 31%.