Joints Blamed in Nimitz Collapse : Disaster: UC engineers find that freeway decks were not adequately joined. Retrofitting of columns might not have prevented the tragedy.

BERKELEY —  A stretch of the double-decked Nimitz Freeway in Oakland collapsed during last week’s earthquake because the top level was not adequately joined to the bottom level, UC Berkeley civil engineers said Wednesday.

The columns supporting the upper roadway of Interstate 880 simply spread out from the bottom like a dancer doing a split, allowing the upper roadway to fall onto the lower lanes. “We’ve never seen a failure like this before,” said engineer Jack P. Moehle.

The finding, announced after a week of investigation by more than 100 UC Berkeley engineers and scientists, contrasts sharply with the prevailing belief that the 35-foot columns simply gave way under the stress of the magnitude 7.1 temblor. It also suggests, according to the researchers, that retrofitting the columns alone, and not the joints, probably would not have prevented the collapse of that section of the Nimitz.

The failure of the Nimitz Freeway has perplexed many of the geologists and engineers who have studied it in the last week. Many potential explanations for the collapse have been offered, ranging from failure of the supporting columns themselves to the existence of a freak resonance between the frequency of the quake’s shock waves and the structure’s own characteristic vibratory frequency.

Now, a consensus seems to be developing that all of these factors may have played a role in the collapse, but that failure of the joints between the upper and lower columns probably was the key element.

“It’s clear that the failure initiates in and progresses from the joints,” said structural engineer Gilbert Hegemier of UC San Diego. The UCSD group had reached this conclusion last week, he added.

Most researchers have dismissed at least two theories: that the double-decked freeway was not built up to standards required by the building code in existence then, and that the structure was inherently unstable. Former state highway engineer Jim See charged earlier this week that engineers had not used as much reinforcing steel as was required by existing codes when the structure was built in the 1950s.

Moehle’s own calculations, he said, indicate that the structure was at least twice as strong as required by the code and that there were no inherent instabilities.

UC Berkeley geotechnical engineer Ray Seed also noted that there was no significant ground failure under the freeway structure that could have caused it to collapse. The deep mud layers under the roadway could have amplified the shaking somewhat, however.

When the Nimitz was built in 1957, its joint construction was standard, the researchers noted. Joints today have considerably greater reinforcement, they said.

At issue is how the columns supporting the upper roadway of the Nimitz are joined to those supporting the bottom one.

At some locations along the one-mile length of the collapsed structure, the upper and lower columns were constructed essentially in one piece, with several steel bars running continuously through both columns. While this provided some measure of strength, Moehle said, there was no horizontal reinforcing steel connecting these columns to the bottom roadway, as required by current building codes. Hence, the top column could shear loose from the rest of the structure, allowing the roadway to collapse.

At other sites along the collapsed roadway section, Moehle said, the joints were even weaker, lacking the vertical steel rods. The joints of those columns were simply unreinforced hinges, with the top columns fastened to the lower ones with a weakly reinforced concrete. Moehle believes that the failure began in these weaker joints and spread to others.

He speculated that the weaker joints were installed where Caltrans planned to construct off-ramps, but never did. Caltrans officials reached by telephone said they had no comment.

Whether the defective joints could have been detected and repaired before the quake is not clear. “The problem is very easy to see with 20-20 hindsight . . . but it’s not clear that an engineer poring over the blueprints (before the quake) would have found it,” Moehle said.

But, added Vitelmo Bertero, director of Berkeley’s Earthquake Engineering Research Center, “had Caltrans spent $200,000 to $300,000 to study the Cypress structure after the 1971 (Sylmar) quake, they could have recognized the problem and developed technology to save the structure. . . . It’s an expensive job, but it could be done.”

Some types of reinforcement might not have been useful, however. Many engineers have observed, for example, that the columns of the Nimitz Freeway are themselves not as strong as they could be. The columns are reinforced only with vertical steel rods; in more modern bridges, the vertical rods are also encircled by steel rods that prevent the vertical rods from expanding outward, allowing the column to “explode.”

The columns on many older bridges are being encased in metal and concrete to prevent such an explosion from occurring. Critics have charged that the columns of the Nimitz should have been reinforced in this manner. But, said Moehle, reinforcing the columns alone would not have prevented the structure’s collapse.

The lack of damage to buildings in the immediate vicinity of the Nimitz has led some researchers, particularly Piotr D. Moncarz of Failure Analysis Associates Inc. of Palo Alto, to speculate that a unique interaction between the Nimitz’s own vibrational frequency and the frequency of the earthquake’s shock waves may have contributed to the failure and collapse.

Moncarz’s scenario is analogous to pushing a child on a swing. If pushes are given randomly, they will have little effect on how high the child swings. But if they are applied precisely when the child is at the end of his upswing, small pushes will have a big effect in increasing the height reached.

Similarly, if each shock wave hit the Nimitz at the precise time it was beginning its motion in the direction of the wave, the waves’ effects would be greatly magnified, accentuating the forces acting on any weak link, such as the joints.

“It seems like this had to have been an important factor,” said civil engineer Nels Roselund of Kariotis & Associates in Pasadena. “It’s certainly a contributing factor,” added UCSD’s Hegemier.

Looking back, Moehle cautioned that Caltrans should not be greatly castigated for failing to recognize and repair the problems on the Nimitz structure. He noted that engineers must balance the risks of every structure against the total cost of repair and reinforcement.

“There are limits to what one can do,” he concluded. “To bring all buildings and bridges up to code would (financially) ruin society.”

The scientists also investigated the collapse of a section of the Bay Bridge. In contrast to the Nimitz Freeway section, the bridge collapsed because its east end was pulled away from the center portion by ground failure on approaches to the bridge. The whole east end of the bridge moved seven inches, UC Berkeley civil engineer Hassan Astaneh said, pulling the span off the six-inch bracket connecting it to the west end and causing it to fall.

Astaneh noted that a series of 24 one-inch bolts holding the collapsed span in place were sheared off by the quake’s force, allowing the span to fall. But, had the bolts not sheared, he said, the pier itself would have been severely damaged and repairs would have taken much longer than the three weeks now envisioned.

NIMITZ FREEWAY COLLAPSE

1. Many experts originally believed freeway columns simply gave way under stress of quake. Now, scientists say the columns cracked at joints to lower deck. Girder reinforcement rods did not extend to upper part of columns, leaving joints vulnerable. 2. Columns dropped down and out, collapsing parts of upper deck. Report says construction techniques used in the joints, unlike many current techniques, were not adequate to withstand the 7.1-magnitude temblor.

STORIES, PICTURES: A3, A30