U.S. Department of Transportation
Federal Highway Administration
1200 New Jersey Avenue, SE
Washington, DC 20590

Skip to content U.S. Department of Transportation/Federal Highway AdministrationU.S. Department of Transportation/Federal Highway Administration

Federal Highway Administration Research and Technology
Coordinating, Developing, and Delivering Highway Transportation Innovations

Public Roads
This magazine is an archived publication and may contain dated technical, contact, and link information.
Public Roads Home | Current Issue | Past Issues | Subscriptions | Article Reprints | Author's Instructions and Article Submissions | Search Public Roads
Publication Number:      Date:  Fall 1996
Issue No: Vol. 60 No. 2
Date: Fall 1996


Aftermath of The Kobe Earthquake

by Hamid Ghasemi, Hisanori Otsuka, James D. Cooper, and Hiroyuki Nakajima

On Jan. 17, 1995, the Hanshin/Awaji Earthquake struck the densely populated Kobe, Japan, area with a Richter magnitude 7.2. In terms of magnitude, this earthquake was not as large as some previous earthquakes in Japan. However, the close proximity of its epicenter to an urban area, along with its severe ground motion and large ground displacement, resulted in more than 5,500 fatalities and extensive damage to lifeline systems (i.e., highways, railways, buried pipelines, etc.)

The Japan Meteorological Agency (JMA) reported that the depth of the hypocenter was 14 kilometers, which is considered to be a relatively shallow earthquake. Several different public and government agencies, including the JMA, recorded the peak horizontal ground accelerations during the earthquake. (See figure 1.) These records revealed a powerful ground motion depending on ground conditions and epicentral distance. At solid ground near the JMA Kobe station in particular, the extremely strong ground motion reached its maximum acceleration of 818 gal in the north-south component. Gal is the same as centimeters per second squared (cm/s2), and 980 gal equals 1 g (gravity force). Furthermore, the maximum velocity and maximum displacement of the earthquake motion recorded at this station were 90 cm/s and 21 cm, respectively. The same measurements for the soft ground near the Higashi Kobe Bridge were 91cm/s and 49 cm, respectively. These are extremely high values compared to records obtained in other earthquakes.

Figure 1 - Peak horizontal ground acceleration (980 gal = 1g) recorded over earthquake area by several different public and government agencies.

Figure 1 - Peak horizontal ground acceleration (980 gal = 1g) recorded over earthquake area by several different public and government agencies.

Historically, the Kobe area has been less seismically active than some other parts of Japan. The last major earthquake in this area occurred in 1916 with a magnitude 6.1 at almost the same epicentral location as the 1995 earthquake. There have been 13 major earthquakes in Japan since 1900. Of the 13, this earthquake produced the second largest death toll -- second only to the Kanto Earthquake, which killed more than 140,000 people in 1923. The main reason for the occurrence of such large earthquakes in Japan is the tectonic activity of the Eurasian, the Philippine, the Pacific Ocean, and the North American plates, which surround Japan. The relative movement of these plates is the main source of strain build-up in the crust of western Japan.

The earthquake caused the collapse of an 18-span viaduct section of the Hanshin expressway. The earthquake caused the collapse of an 18-span viaduct section of the Hanshin expressway.

Today, Kobe is still recovering. Housing remains a significant problem. Traffic is being rerouted along several detours; the Hanshin Expressway, the major transportation route between Osaka and Kobe, is still being repaired and/or rebuilt to more stringent seismic requirements. Some buildings have been demolished and are under reconstruction, while many others are being repaired. Most existing highway bridges are being strengthened using various retrofitting schemes, e.g., steel column jacketing and seismic isolation bearings.

The Japanese began designing bridges for seismic loads after the great 1923 Kanto Earthquake. Since their initiation, the seismic regulations have been amended several times. At first, the only seismic design provision for steel and concrete bridges was the application of a lateral design force equal to 20 percent of the weight of the structure. In 1971, Japan's Ministry of Construction issued the first guide specifications for the seismic design of highway bridges. According to the guide specifications, the Japanese designed their bridges based on a seismic coefficient method using allowable stress. The seismic lateral design force was then determined according to seismic zone, importance, structural response, and site condition. The latest seismic design specifications were issued in 1990. A major change was a required check for the ultimate capacity of reinforced concrete piers using a ductility approach under an assumed magnitude 8 earthquake, where force levels of up to 1.0 g are applied to piers.

The new reinforced concrete piers for the 18-span section of the Hanshin Expressway that was destroyed in the earthquake.The new reinforced concrete piers for the 18-span section of the Hanshin Expressway that was destroyed in the earthquake.

The lessons learned from the Hanshin/Awaji Earthquake have prompted a necessary revision of the current seismic design standards (1990 Guide Specifications) for highway bridges in Japan. Since these revisions are not final, a tentative guide specification is currently being used for the seismic design of new highway bridges and seismic strengthening of existing ones as an emergency measure. The Hanshin/Awaji Earthquake was the first experience with high accelerations for engineers in Japan. This will have a profound effect on Japanese standards for seismic design of structures, just as there were major revisions to the California Department of Transportation (Caltrans) design specifications after the 1971 San Fernando Earthquake.

Figure 2 - Comparasion of 1990 design spectra and 1995 tentative design spectra, which is based on the strong acceleration of the Hanshin/Awaji Earthquake.According to the 1995 tentative design specification, the ultimate capacity of piers must be checked for force levels of up to 2.0 g with a tentative design spectra based on the strong accelerations of the Hanshin/Awaji Earthquake recorded by JMA. The new spectrum resembles the Acceleration Response Spectrum (ARS) design curves of Caltrans, which indicates that Japan and Caltrans in some ways are moving toward one another in designing highway bridges. The tentative design spectra reflect the reversal of the relative response of different soil types and a rapid increase of standard seismic coefficients compared to Japan's 1990 design spectra. (See figure 2.)

Highway No. 3, Kobe Line

The most heavily damaged section of the Hanshin Expressway network, the 27.7-kilometer (km) Highway No. 3, Kobe Line, is being restored and is scheduled to open to traffic by the end of 1996. The restoration includes strengthening existing bridge piers with steel column-jacketing and reconstructing an 18-span viaduct section that collapsed due to shear and flexural failure of columns. The failure was triggered at the point where some of the longitudinal reinforcement was terminated. The 180 longitudinal bars at the base of the 3.1-meter (m)-diameter columns were reduced to 120 bars 2.5 m above ground level. These columns were designed according to pre-1980 design specifications and were scheduled for retrofitting. The 1995 tentative design specification no longer permits the termination of longitudinal reinforcement in columns at mid-height. It also calls for reduction of shear reinforcement spacing from a maximum of 30 cm to 15 cm.

Reconstruction of the piers of the 18-span viaduct section was completed last February. The new reinforced concrete piers are about 3 m wider than the old piers in the transverse direction. This has resulted in a reduction of one lane in each direction on National Route 43 beneath the Kobe Line. The piers' longitudinal bars are continuous and have shear reinforcement at 12.5-cm spacing through the entire height. Two 9-span, steel box girder superstructures with a steel deck will be used to reduce dead weight. The structure uses Menshin bearings (Japanese terminology for seismic isolation bearings) to distribute forces and dissipate energy. The two lead rubber bearings per pier are blocks measuring 1100 by 1100 by 237 millimeters (mm) and contain four lead cores, each with a diameter of 14 cm.


The typical foundation of Highway No. 3 consists of a footing on cast-in-place concrete piles. Virtually every column-footing connection was investigated after the quake by direct observation at the time of excavation. Very few cracks were found on the top surface of footings. As for reinforced concrete piles, bore-hole cameras were used to determine their condition. Tension cracks were observed near the top of piles supporting the piers of the 18-span viaduct section that collapsed. These piles were repaired by drilling a 36-mm-diameter hole at their center and subsequently filling it with very fine cement slurry.

Isolation Bearings

Most superstructure damage was caused by the brittle failure of the steel bearings. The percentage of steel bearings that suffered from severe damage was much higher than that experienced by rubber bearings. Reports from an early investigation by the U.S.-Japan Program in Natural Resources (UJNR) indicated that premature failure of some bearings appears to have reduced the seismic loads in their supporting substructures by uncoupling the superstructure from the violent ground motions. This fuse-like action may have saved the columns from shear and flexural failure and saved a number of spans from collapse. This points to the potential effectiveness of seismic isolation. Today, the Japanese are relying more on the use of seismic isolation bearings. In Kobe, vulnerable steel bearings are being replaced by these newer bearings, which have proven to be effective in reducing seismic forces in buildings.

With this approach, bearings are placed between the superstructure and the top of the substructure (piers and/or abutments). Under normal conditions, they behave like regular bearings. However, in the event of a strong earthquake, they add flexibility to the structure by elongating its period and dissipating energy. This permits the superstructure to oscillate at a lower frequency than its piers, which results in reduction or elimination of deformation of the substructure components beyond their elastic range, particularly at locations that are difficult to inspect or repair (e.g., piles). Isolation bearings are very effective at sites with stable soil. Approximately 90 U.S. bridges have incorporated this technology since 1985.

Today in Kobe, Japanese engineers are innovative. They are reconstructing the Kobe Line's Ben Ten segment, a 19-span continuous bridge with a seismic isolation bearing at the base of each column. This approach was adopted since the rigid connection of footing and columns would have made the bearing capacity of the existing piles inadequate for the new seismic loads. The bridge has a steel rigid frame substructure and steel box girder superstructure. This technique was used for the first time in Japan.

Before the quake, the Matsunohama Bridge, located 39 km from the epicenter, was the only isolated and instrumented bridge in the Osaka-Kobe area. The bridge has four continuous spans on single-column reinforced concrete piers, a steel box girder superstructure, and concrete deck. Two lead rubber bearings are used in each pier. The recorded acceleration in the longitudinal direction at one pier was 0.19 g at the girder, 0.20 g at the top of the pier, and 0.15 g at ground level. It is clear that the ground acceleration was not profound enough to test the effectiveness of the isolation bearings. However, few buildings were protected by isolation bearings.

One protected building houses the Technical Research Institute of the Matsumura Gumi Corporation. It is isolated with high damping rubber bearings installed between the footing and columns. The building is a three-story, reinforced concrete, rigid-frame structure that is separated from a fixed-base steel office building through expansion joints. A comparison with the fixed building reveals that the peak floor accelerations were reduced at the roof by a factor of approximately 4.9 in the north-south direction and 2.5 in the east-west direction.

Figure 3 - Diagram of the full-height circular steel jacket used to retrofit the refinforced concrete bridge column on the Hanshin Expressway.

Retrofitting Reinforced Concrete Columns

Since the earthquake, the Public Works Research Institute (PWRI) of the Japanese Ministry of Construction has been conducting a series of experimental lateral load tests at its research facility on the performance of different retrofit techniques for strengthening bridge piers designed before 1980. These piers are inherently inadequate in flexural strength, flexural ductility, and shear strength.

The standard retrofitting procedure for reinforced concrete columns in Japan is to increase the cross section of columns by the addition of reinforced concrete, which is then surrounded by steel jackets either at locations of premature termination of main reinforcement or over the column's full height. The function of the additional reinforced concrete is to increase the flexural strength of the column. Steel jacketing is the external encasement of columns using prefabricated steel shells that are welded together on site. If the road clearance is insufficient and it is impossible to increase the column's cross section with the additional reinforced concrete, steel jacketing alone is then used to increase the flexural ductility and the shear strength of the column.

Most existing bridge columns on the Hanshin Expressway network in Kobe are being retrofitted with full-height circular steel jackets. The jacket extends below ground level and terminates 50 mm above the footing. Longitudinal reinforcement is added outside of the steel jacket. Epoxy grout is used to anchor these bars in holes drilled into the footing. The anchorage length is about 30 diameters. Shear studs of 22-mm diameter and 130-mm length are welded around the steel jacket at a specified spacing to transfer earthquake force from the longitudinal reinforcement to the steel jacket. The reinforcement and shear studs are protected by the concrete jacket that extends from the top of the footing to ground level. The gap between the column and steel jacket is filled with epoxy resin. (See figure 3.)

Steel jacketing was originally developed by the University of California at San Diego for retrofitting circular columns. In California practice, circular and elliptical steel jackets are used to enhance the ductility of circular and rectangular columns, as well as to inhibit shear failure caused by an inadequate lap splice of longitudinal bars. (The Japanese retrofit rectangular and square columns with rectangular and square jackets, respectively.) The steel jackets are full length with a 50-mm space between the jacket and the footing or cap beam to prevent bearing of the jacket on either footing or superstructure due to lateral displacement of the column, which may increase the moment demand. Pure cement-based grout is used to fill the gap between column and steel jacket to ensure composite action between the jacket and the column.

Retrofitting Steel Columns

The Hanshin/Awaji Earthquake also showed the vulnerability of hollow steel piers to local buckling, which greatly reduced its bearing capacity in the vertical direction. Two rectangular steel piers were also crushed in compression due to the rupture of the welding at the corners under the high level of horizontal ground acceleration.

Tests have been initiated by PWRI in cooperation with other agencies to determine the performance of both existing and newly constructed rectangular and circular steel bridge piers. Different stiffening techniques for strengthening under a constant compressive axial load and cyclic lateral loads are being investigated. Results have shown that some of the new techniques for the strengthening of rectangular steel piers with inside angle and outside corner plates or inside angle plates will increase ductility as well as prevent cracking at corners after local buckling has begun in both rectangular and circular steel columns. (See figure 4.) For circular steel columns, an additional thin layer of steel confinement is provided at the bottom of the pier to restrain the progress of out-of-plane deformation after local buckling. However, in the mean time, the tentative design specifications suggest filling steel bridge piers with concrete for reconstruction and repair. This approach has been used on many steel bridge piers in Kobe.

Conclusions and Lessons Learned

At least 60 percent of all bridge structures in the Kobe area were damaged. The damage to the majority of reinforced concrete columns was no surprise once the column reinforcement details were examined. The predominate failure mechanism of these columns was either shear failure or flexural shear degradation failure. These two types of failure are related to a nonductile design that is the result of inadequate shear or confinement reinforcement. These bridges were designed and constructed in the 1960s before the introduction of modern seismic codes.

Before the Hanshin/Awaji Earthquake, the Kobe area was considered to be a region of low or even negligible risk of moderate earthquakes even though a past earthquake hit the area causing light damage.

Both of the observations are applicable to the United States, especially to the states east of the Rocky Mountains that are within the seismically active New Madrid Zone. According to seismologists, this fault is capable of generating moderate to severe earthquakes in the next 100 years. The bridges in the central and eastern United States are very similar to the types of bridges in Japan. They are mostly steel-beam superstructures supported by steel bearings and foundations that were designed with no consideration of seismic loads.

The question is: Are we prepared? Can our bridges survive a Kobe-type event? It is very important to increase public education and awareness, especially in areas where damaging earthquakes are infrequent. States must recognize and identify their vulnerable structures. Due to limited funds, these structures must be prioritized and those that are most important to the community must be retrofitted first.

The Hanshin/Awaji Earthquake showed that very large bearing seats and confinement for reinforced concrete columns can prevent catastrophic collapse. In addition, the observation that premature failure of steel bearings may have allowed the pier to move freely under the superstructure during the earthquake, thus significantly reducing the pier column forces, points out the potential effectiveness of seismic isolation.

Seismic isolation is an attractive alternative when other retrofitting measures lead to undesirable substructure performance. The approach has been used successfully in Italy, Japan, and New Zealand for building and bridge structures. More recently, the United States has begun to implement isolation technology in bridges.

In recent years, many seismic isolation and energy dissipation systems have been proposed to the bridge industry in the United States by different manufacturers. The proprietary nature of these systems and the lack of knowledge of their long-term performance have made them less desirable to many engineers.

However, for the past two years, the Highway Innovative Technology Evaluation Center (HITEC) has been collaborating with Caltrans and the Federal Highway Administration (FHWA) in developing a test and evaluation program for isolation systems proposed for both retrofit and new construction bridge applications. The program will provide verifiable, credible information on the functional performance (seismic and nonseismic), constructibility/practicality, durability, materials characterization, and dynamic behavior of the various systems and components submitted for evaluation. The program is unique for its full-scale dynamic testing of all the systems.

The tests are being conducted at the Department of Energy's Energy Technology Evaluation Center in California. The test plans were developed under the guidance of bridge engineers from the departments of transportation of five states, as well as experts and researchers from industry and academia. Thirteen manufacturers domestic and international with a total of 14 systems have participated in this program. FHWA is funding the program under the Applied Research and Technology grant program. The results of HITEC tests will help highway agencies and bridge designers in the United States identify the appropriate systems for retrofitting existing bridges or for incorporation in new bridges.

It is clear that the existing bridges designed to preseismic codes must be upgraded if public safety is to be improved and Kobe-type disasters are to be avoided. Bridge engineers have made considerable progress in recent years, and it is clear that current knowledge of structural behavior and seismic hazards has greatly improved the performance of modern structures under seismic forces.

Dr. Hamid Ghasemi is an FHWA research structural engineer. At the invitation of PWRI, he spent one month studying isolation bearing technology at the research laboratories in Tsukuba, Japan.

Dr. Hisanori Otsuka is the head of PWRI's Earthquake Engineering Division.

James D. Cooper is the chief of the Structures Division of FHWA's Office of Engineering Research and Development.

Hiroyuki Nakajima is the senior engineer of the Hanshin Expressway Public Corporation.



Federal Highway Administration | 1200 New Jersey Avenue, SE | Washington, DC 20590 | 202-366-4000
Turner-Fairbank Highway Research Center | 6300 Georgetown Pike | McLean, VA | 22101