Weld Metal Tests Stir Steel Building Concerns

Engineers do not allow this material to be used for building bridges or oil pipelines. They know it’s not strong enough.

But for the steel-frame structures that we live and work in, they have allowed builders to use it by the ton.

Obscured beneath eye-pleasing finishings, the material is a weld metal that fuses thick beams, columns and plates of structural steel. Walk through most any steel building from San Diego to Seattle and it surrounds you.

For more than 20 years, builders have used the weld metal, known as “120” or “E70T-4,” with confidence. The assumption has been that the welds help make the steel structures flexible and sturdy enough to withstand even a major earthquake.

But an array of university test results--the first of which were shared with steel construction insiders within months of the January 1994 Northridge earthquake--show the assumptions were based on a faulty premise:

Welds made with E70T-4, it turns out, are not capable of performing as had been expected.

Leading researchers now say that the West’s most widely used weld metal is so prone to brittle fracturing that it should be banned from structural uses. Their tests show that E70T-4 has just one-fourth the resistance to fracturing of other available materials.

Consistent with its performance in the Northridge earthquake--when thick, welded connections of steel fractured--the metal has failed in laboratory tests at stress levels far below what would be expected in a major seismic event.

“You should never use this kind of material,” said John W. Fisher, a professor of civil engineering at Lehigh University, in Bethlehem, Pa., whose research has been supported by grants from the federal government and by the steel industry. “We should ban its use . . . I think that there’s great risk that it’s going to cause fracture.”

Michael D. Engelhardt, an engineer at the University of Texas at Austin, confirmed in May 1994 that the weld metal fractured under only moderate stress.

“There’s just a consistent picture here of real poor performance,” he said in an interview. “. . . we now realize how important it is to have high-toughness weld metal in these connections . . . once you see it, it’s fairly obvious.”

Yet what appears obvious on campus has not gained easy acceptance elsewhere. With the encouragement of the leading distributor of E70T-4, authorities have proved reluctant to prohibit the material.

L.A. Bans Material

Building officials in Los Angeles finally banned use of E70T-4 and other metals with low fracture resistance in July for new construction. And the distributor, Lincoln Electric Co., began advising customers as of June that weld metals it sells with higher resistance than E70T-4 “should be considered” for “seismic applications.”

But E70T-4 remains legal to use elsewhere in California and throughout the earthquake-prone West. Even in Los Angeles, officials have no plan to force the retrofitting of the hundreds of buildings that are believed to be dependent on the fracture-prone weld material. Most of these structures, well away from the Northridge epicenter, have not been subjected to mandatory, quake-related inspections.

Richard Holguin, chief of the city’s engineering bureau, said he is now convinced that E70T-4 “fractures very, very abruptly.”

The city’s new policy bars the use of E70T-4 and other fracture-prone metals in major connections. The materials were earlier banned for use in earthquake repairs.

Citing the costs involved, Holguin said it would be impractical to require buttressing of the connections welded with E70T-4 that were not necessarily damaged by the earthquake. Holguin said he assumes most of the 1,500 steel structures in the city have been welded with E70T-4.

At Lincoln Electric, representatives said the Cleveland-based company has acted responsibly and has sought only fair treatment of its entire “Innershield” line of welding metals, including E70T-4. They said that further research is still needed to determine the extent to which the weld metal itself, or other factors, caused damage to the connections.

“There is still not a unanimous opinion within the engineering community,” said the company’s welding design engineer, Duane K. Miller, adding: “Until it’s better defined, we’re accepting the fact that [new] requirements should be adhered to.”

Quick Application

Because E70T-4 can be applied faster and at higher volumes than other available weld metals, major steel fabrication firms have preferred using it. The weld metals that are more fracture-resistant are not as conducive to rapid production, due to their narrower diameters and the lower temperatures at which they are applied.

An executive of Herrick Corp.--the West’s largest fabricator of steel--acknowledged that E70T-4 has been the firm’s material of choice for the past 15 years.

This fast, “high deposition” welding metal lowers costs by boosting welders’ productivity--a crucial consideration for fabricators and for building owners.

Herrick’s vice president, Roger E. Ferch, said the company has used E70T-4 with the full approval of structural engineers. He also said Herrick experienced a “20 to 30% reduction in production” last year when an engineer ordered the firm to stop using E70T-4 on a hospital project after repeated problems with cracking.

Other forms of construction have not been without problems, either. The failure of unreinforced masonry contributed to scores of deaths in the 1933 Long Beach earthquake. In the Northridge quake, three- to four-story wooden apartment buildings--often with ground-level parking garages--buckled and in one case collapsed, killing 16 people. Some concrete structures, built before 1970 with inadequate flexibility, also fared poorly.

Stunning Findings

The realization that steel buildings also are vulnerable, however, has stunned building officials and engineers.

The Northridge quake, which struck at 4:31 a.m. on Jan. 17, 1994, ruptured welded connections in at least 150 structures. No deaths resulted from ruptures and no steel buildings collapsed, but some structures were damaged extensively. In Santa Clarita, a two-story building, owned by an insurance company, was so crippled that its owners chose to demolish it.

There is no certain explanation of why steel buildings were not hit even harder--or why similar damage wasn’t inflicted by earlier earthquakes. Some experts note the structures’ “redundant” designs, which allowed them to stay standing while sustaining internal damage. Others note that the most severe ground shaking from both Northridge and the 1971 Sylmar earthquake occurred away from downtown Los Angeles, where many of the tallest steel buildings are located.

As is typical in construction when misfortune calls, various players involved with the building of the steel structures that were damaged in the Northridge quake have not hesitated to cite factors apart from their own efforts:

Representatives of Lincoln, the leading seller of E70T-4, say that the product was sound--but not used under strict-enough controls. Steel fabricators point to the errors of structural engineers’ designs. Structural engineers, many of them admittedly unschooled when it comes to welding, blame poor workmanship by the fabricators or poor inspection.

Indeed, researchers continue to examine the role played by factors other than the weld metal itself. The geometry and individual details of beam-to-column connections, the strengths of the steel beams and columns, welding workmanship and vigilant inspection all are important.

But tests show that not even ideal construction conditions can overcome the inherent vulnerability of E70T-4.

“I don’t think it’s possible to have a tough [fracture-resistant] weld with this material,” said Fisher of Lehigh University. On the other hand, Fisher said, engineers have gotten satisfactory results when they tested similarly configured connections made with higher-quality weld metals.

After the Lehigh engineers published their research, Fisher agreed to assist owners of a damaged building in Santa Monica who are suing Lincoln Electric, claiming that E70T-4 is a defective product; Lincoln denies the allegation.

Lincoln’s engineer, Miller, said he believes the conclusion reached at Lehigh regarding E70T-4 is “fundamentally flawed” and based on imprecise tests.

John F. Hall, an associate professor of civil engineering at Caltech who has studied the performances of structures in the Northridge quake, said the laboratory-confirmed weakness of E70T-4 is consistent with the brittle fractures he has observed.

Hall first presented his concerns about the vulnerability of steel buildings in spring 1994, saying that a magnitude-7.0 earthquake--roughly twice the intensity of the 6.7 Northridge temblor--could level some structures in Los Angeles or the San Francisco Bay Area.

“Our nearby faults are capable of producing a strong enough earthquake that could cause collapses of steel buildings,” Hall said in an interview.

According to engineers and inspectors who have studied the Northridge damage, the fractures of steel connections typically initiated within the weld metal itself. Many of the weld cracks spread, tearing into and clear through adjoining columns or beams.

To the extent the fractured weld metal could be verified, the specialists said, it was E70T-4. And, people familiar with the construction agree that most of the steel buildings erected in the region since the early 1970s have been welded with E70T-4.

“The process of choice for West Coast contractors is self-shielded flux cored arc welding, and the electrode typically used is E70T-4,” Miller wrote, in an article published in 1994 by Lincoln Electric. Based on Lincoln’s worldwide sales of those “Innershield” products in 1993, a company lawyer said that E70T-4 accounted for 11.4% of purchases.

Now, 2 1/2 years after Engelhardt’s findings at the University of Texas, E70T-4 remains legal to use in most areas. Among those whom Engelhardt said were privy to his research, as of May 1994, were representatives of Lincoln Electric and two major steel fabricators based in California, Herrick Corp. and PDM Strocal Inc.

Lincoln Electric’s Miller said the research conducted by Engelhardt was not definitive and that there was no reason, as of mid-1994, for the company to advise customers that E70T-4 was prone to fracture. Engelhardt, for his part, said he believes the quality of welding workmanship, design details and the resistance of the weld metal all are important.

Lawsuit Filed

The two top executives at PDM Strocal, Fred and David Long, said they stopped using E70T-4 several years ago, for reasons they declined to discuss. Records show the company was sued, in July 1996, by USC over extensive weld cracking found before and after the Northridge quake at the eight-story Kenneth Norris Jr. Comprehensive Cancer Center, built east of downtown near County-USC hospital.

According to a July 1994 report by the architectural firm of Anshem + Allen, 33% of 160 significant welds at the cancer center that were made with E70T-4 cracked. The firm at that time ordered the remaining unfixed welds to be repaired or dug out of the structure and replaced.

“We find that [brace frame] welds made with NS-3M [the weld metal’s trade name] are defective,” the firm’s report said, adding: “The testing program indicated the NS-3M 120 electrode field welds do not have sufficient strength and ductility to fulfill their functions at [brace frame] joints in a major earthquake.”

PDM Strocal--which, according to Los Angeles Superior Court records, stopped using E70T-4 on the cancer center as of May 1993 because of the cracking problems--has defended its work as proper. Last March, PDM sued USC and the general contractor, seeking payment for its work. PDM alleges the weld fractures were caused entirely by the architectural firm’s design of the connections; the architects maintain their design was adequate.

Lincoln Electric has defended the performance of E70T-4 that it supplied to PDM for use in the Norris Cancer Center--attributing the cracking problems to the design. The company is not a defendant in the litigation.

At the cancer center and elsewhere, Lincoln saw potential trouble for E70T-4 and its other weld metals as the damage to steel structures emerged after the Northridge earthquake. Interviews and records obtained by The Times show that Lincoln began working soon after the quake to persuade authorities not to prohibit use of its materials.

In spring 1994 reports to his bosses, Miller said he had dissuaded authorities and structural engineers in the West from immediately halting the use of Lincoln’s “flux-cored” welding products, which included E70T-4. Miller wrote that his offers of advice were well-received by Holguin, then assistant chief of the city of Los Angeles’s Building Bureau.

“I offered to review any specifications they were about to release, and he [Holguin] was eager to receive our input along this line,” Miller wrote, on April 19, 1994, to Lincoln’s vice president of business development. “This opportunity, I believe, is highly significant. It would enable us to challenge provisions that would be unnecessary and/or harmful to our position and to that of our customers.”

Miller also warned that if Lincoln did not take steps to achieve quality workmanship for connections that would be tested at the University of Texas, “it is apparent to me that this product [E70T-4] will be banned from the marketplace.”

Referring in another April 1994 memo to his contacts in California with “various authorities” and university researchers, Miller said he had sought to “minimize the negative impact that could potentially have occurred from this earthquake,” adding:

“The fact that self-shielded flux cored electrodes have not been banned is evidence that we are on the right path. Had we not been present, I am confident that this is one of the actions that would have been taken.”

Two months later, Miller warned the company vice president of the need to “circumvent unnecessary code provisions” being proposed by other local governments in California and by the state government. Miller also wrote in that memo, dated June 22, 1994: “Relationships must be developed with the state offices just as they have been with the L.A. City Hall in order to preempt the issuance of unnecessary, costly and unjustified provisions.”

Fought ‘Simplistic’ Fix

In recent interviews, Miller said that Lincoln Electric’s main objective in the wake of the earthquake was to guard against regulations that might be based on incomplete evidence. Miller said he wanted to protect the “family” of Lincoln’s welding products--including E70T-4 and other weld metals with both low and high fracture resistances.

“Our primary goal from the start of this was to look for a reliable solution to this problem of fractured connections,” Miller said. “We did not want to permit people to jump to simplistic solutions.”

Holguin, for his part, said that he did confer with Miller, but exercised independent judgment.

“They’re basically in the business of selling weld metal,” Holguin said. “So you always take what they say with a grain of salt.”

At Herrick, the top West Coast fabricator, company Vice President Ferch defended the company’s continued use of E70T-4, saying it has always been done with the approval of structural engineers. Ferch also acknowledged that he knew, as of June 1994, that E70T-4 performed poorly in the testing the company was participating in with the University of Texas.

Since 1995, he said, Herrick has not used E70T-4 for what he termed earthquake-sensitive welds.

“That’s the wire that we have used in our shops, probably for the last 15 years, on all projects, unless a special requirement was there,” Ferch said, in an interview in April. “That’s a standard electrode that we have used.”

Records and interviews show that among the major projects for which Herrick has used E70T-4 since the Northridge quake are:

* The 28-story Gateway transit building in Los Angeles, adjacent to Union Station. The building was welded from September 1993 through July 1994.

* A regional hospital complex near San Bernardino, welded with the material from spring 1995 until October 1995--when a project engineer banned it, citing repeated cracking problems. The complex, still under construction, borders two significant faults capable of at least a magnitude-7 earthquake--the San Andreas and the San Jacinto.

* A 30-story office tower in Honolulu, the tallest building constructed in the Hawaiian Islands. Owned by First Hawaiian Bank, the project was welded from early 1994 through late 1995.

Ferch and another Herrick executive did not return calls placed to them this month. Referring last spring to the San Bernardino hospital project, Ferch said:

“You’ve got a product that meets the [engineer’s] plans and specifications, that is fully supported by a rather large company, called Lincoln Electric. We are not talking seismically sensitive welds.”

The private engineer responsible for the hospital complex, Jefferson W. Asher, said that, as a precaution, he would not again allow use of E70T-4 for those welds. But Asher said he is satisfied the hospital will be safe.

At the Gateway building in Los Angeles, now home to the Metropolitan Transportation Authority, E70T-4 was used for hundreds of major, “full-penetration” welds during 1993 and 1994. The project’s structural engineer, John A. Martin Jr., said the most significant connections were reinforced before the building was completed.

Martin said he no longer allows the material to be used for load-bearing welds on his jobs. Still, he said he sees the risk for the Gateway building and other structures welded with E70T-4 as one of potential property damage cost, not a threat to life.

Honolulu Construction

After the earthquake, Holguin said city engineers unsuccessfully recommended to Martin’s firm that steel cover plates be installed at the Gateway building to further strengthen the connections welded with E70T-4.

“Legally, quite frankly, we didn’t have any ability to force them to change their design,” Holguin said.

In Honolulu, the structural engineer for the First Hawaiian office tower, Gary Chock, confirmed that the material was used. Chock said the structure is not in an area of high seismic risk and that he is satisfied it is safe. He said he followed standard practice by allowing Herrick to select the welding materials that were used.

After learning of weld defects that were emerging during fabrication, Chock said, he took steps to ensure that they were located and repaired as the building was erected. According to the developer, 90 tons of weld metal went into the structure.

The sustained use of E70T-4 is eye-opening to engineers who have overseen the university research.

“I am surprised that it’s being used, with what we know today,” said Fisher, who is director of the National Science Foundation Engineering Research Center at Lehigh. “Engineers who use it, knowing what has happened, put themselves at risk.” Fisher said he would not use E70T-4 anywhere, regardless of an area’s susceptibility to earthquakes.

Fisher said he first publicly shared his concerns at a September 1994 conference in Los Angeles sponsored by the Structural Engineers Assn. of California. “It was pretty clear in our view by then that the primary cause of the problem was brittle weld metal,” Fisher said, referring to his colleagues at Lehigh.

Fisher’s research has included the study of 13 steel connections from five separate buildings that were damaged by the Northridge earthquake. The work has been supported by a grant from the National Institute of Standards and Technology.

Still, most authorities have not prohibited use of E70T-4 and other fracture-vulnerable weld metals. The lack of action runs counter to a recommendation made in August 1995 by a blue-ribbon research group whose work has been backed by $10.6 million in grants from the Federal Emergency Management Agency.

Composed of representatives from structural engineering firms and various universities in the state, the group recommended that only highly fracture-resistant weld metal should be used in “critical joints.”

The weld metal in those applications should be able to absorb 20 “foot pounds” of impact at 0 degrees Fahrenheit, the group advised. By comparison, E70T-4 typically can withstand up to just 5 foot pounds before it fractures.

Welding Group’s Stand

Why haven’t officials moved more swiftly to actually ban the structural use of E70T-4 or other weld metals with similarly low fracture resistance?

Some engineers and inspectors point to the pressures from building owners to keep costs low. Another explanation is that the American Welding Society--whose code book specifies requirements for structural welding in the United States--has declined to oppose the use of E70T-4.

With rare exception in steel construction across the United States, the AWS code rules. “For structural welding, we go with AWS,” said Tony Re, a building official for the city of Portland, Ore.

The AWS code and the society’s position statements are particularly significant because many officials, and even structural engineers, lack knowledge or are uncertain regarding welded construction. Often, The Times has found, they defer to contractors and steel fabricators for selecting weld metals.

Even in Anchorage, Alaska--where a magnitude-8.4 earthquake devastated the city in 1964--federal engineers said they relied on the AWS code in allowing the use late last year of a weld metal with low fracture resistance, comparable to E70T-4. Repairs were made this year to scores of cracks found in welds provided for a two-story hospital being built at Elmendorf Air Force Base, according to project engineers. The weld metal applied, known as E70T-7, is used widely on the East Coast.

Because of the cracking problems, said Ray C. Decker, of the Army Corps of Engineers, “We may have some recommendations for changing the [weld metal] criteria” for future projects.

Warren G. Alexander, a civil engineer who, as an official of New York state prevented use of E70T-4 in bridges and in tall-building construction during the past 20 years, has urged the welding society to take a stronger stand.

“This welding product has been known to have low ductility and toughness for years,” Alexander said in a letter four months ago to the chairman of the society’s Structural Welding Committee. “To argue that toughness is not important is not a good position to take at this point. It’s like defending the safety of cigarettes.”

From the time questions about weld performance began emerging after the Northridge earthquake, the welding society’s lead voice on the issue has been senior staff engineer Hardy H. Campbell III. In correspondence and published articles, he has defended the welds--while criticizing California’s structural engineers for not grasping how their designs can cause connections to fail.

In letters sent in August 1994 to colleagues of the welding society’s “Presidential Task Group,” Campbell warned against code changes.

“Despite the apparent momentum to make special seismic revisions to the [Structural Welding] Code, this should be resisted. The fracture phenomena has very little to do with welding, in point of fact,” Campbell wrote on Aug. 12, 1994, adding:

“I talk with California engineers all the time, and their confusion is exacerbated by the misinformation running rampant. . . . It remains to be seen how effectively the cows can be herded back into the barn.”

Campbell has been well-positioned to influence the conversation:

He and Lincoln Electric’s Miller founded the welding society’s Presidential Task Group, which considered whether to recommend code changes because of Northridge. And as secretary of the AWS Structural Welding Committee, Campbell’s articles have been published nationally.

Positions Criticized

The positions taken by Campbell and Miller have been criticized by Raymond H.R. Tide, a veteran Chicago-area structural engineer who also served on the Presidential Task Group and who has urged use of fracture-resistant weld metal.

The task group report, distributed in November 1995 by the welding society, termed the issue of fracture resistance “complex” and in need of further study. It stopped short of recommending fracture-toughness requirements.

“Research should be conducted to determine the required level of notch toughness for weld metal [that is] subject to seismic loading,” the report said, adding that certain toughness levels “should be considered” by structural engineers.

The welding society distributed the task group report as a position statement regarding the problems posed by the Northridge earthquake.

Campbell was unavailable for comment. His boss, Chuck Fassinger, said any positions taken by the AWS or its committees are advisory and are not binding on individual structural engineers. Fassinger said the Miami-based society might eventually distribute separate guidelines for welding in seismically sensitive regions.

As for whether the report should have flatly recommended against E70T-4, Miller, who was chairman of the task group, said individual engineers must make their own decisions.

Tide was not surprised to hear that use of E70T-4 has persisted.

“As time passes, unless somebody really precludes the use of a product, it will creep back in--especially if there is an economic incentive,” Tide said. “E70T-4 is not an acceptable weld metal. Anyone with a minimum knowledge of welding should know that. . . . If there’s a construction accident, and someone is killed and it’s established that E70T-4 is anywhere near it, the ramifications are rather profound.”

Times researcher D’Jamila Salem-Fitzgerald contributed to this report.

Tuesday: A badly damaged building in Santa Monica.

(BEGIN TEXT OF INFOBOX / INFOGRAPHIC)

Fractured Connections

The January 1994 Northridge earthquake ruptured connections in at least 150 steel buildings. Since then, university researchers have concluded that the weld metal used to fuse most of the connections, called E70T-4, has a low resistance to fracture, compared with other available products. The city of Los Angeles in July 1996 banned the use of this and other fracture-prone weld metals for use in construction. In addition to the weld metal, other factors in determining adequate structural performance are the geometry and overall design of the connection, welding workmanship and inspection.

Spreading Crack

Engineers and inspectors who have examined the damage from the earthquake believe that it began, typically, when the E70T-4 weld metal fractured and the resulting crack then spread into or through an adjoining vertical column or horizontal beam.

Source: Los Angeles Times. Researched by DAVID WILLMAN / Los Angeles Times