On January 17, 1995 at approximately 5:47 am local time, a 7.2 Magnitude (Japan Meteorlogical Agency Scale) Earthquake decimated Kobe, Japan and the surrounding region. An entire order of magnitude stronger than the Northridge Earthquake, the Kobe Earthquake (Hyogo Ken Nanbu Earthquake) lasted for nearly 15 seconds claiming the lives of over 5,000 people, injuring more than 26,000 others, displacing 300,000 residents, destroying over 150,000 buildings, and accumulating an estimated 20 trillion Yen (200 Billion US Dollars) in economic losses (Akai, 1995). The epicenter was located at the northern part of Awaji Island, just 20 km from the densely developed urban city of Kobe, which is home to nearly 1.5 million people (Akai, 1995). With extensive structural damage to ports, harbors, dams, levees, underground transit systems, highways, bridges, and buildings, the aftermath of the Kobe Japan Earthquake was a living laboratory for engineers to explore and learn from. The structural damage depicted in Figure 1 was widespread throughout Japan. Although a wide array of failure mechanisms were identified throughout the region, the primary design flaws include the following:
Excessive column tie spacing
inadequate base plate detailing
improper design emphasis on stength as opposed to ductility
improper hooks
Vertical discontinuity in building stiffness
insufficient drift capacity in RC columns
failure to address effects of soil liquefaction and lateral displacement of soil
combination of heavy roof construciton with lightly braced wood frame construction
Seismology
Kobe Japan is located in an earthquake prone region dominated by complex faulting systems (Akai, 1995). The Hanshin region and Awaji Island are affected by the subducting Philippine Sea Plate and the Plastic Plate. The 1995 Kobe Japan earthquake was initiated by the release of accumulated stress generated by the subsiding, uplifting, and tilting of
tectonic plates. Specifically, the earthquake source can be traced to three tectonic subevents that occurred along the Nojima Fault, which is directly beneath the Akashi Strait. These three events included two strike slip ruptures and a bilateral rupture. Japan Meteorological Agency reported a registered magnitude of 7.2. The epicenter was located at the northern part of Awaji island, just 20 km from the densely urbanized city of Kobe, Japan (Kitigawa, 2004). As shown in Figure 2, the northern coast of Japan experienced the greatest ground shaking. With a 16 km focal depth and a 15 second duration, the earthquake caused 1.7 m lateral surface displacement and 1.0 m vertical ground displacement. Peak ground shaking was .5g, but a limited number of locations experienced peak ground motion that exceeded .8g (Chung, 1996). The overall rupture length was approximately 30 to 50 kilometers (Dickenson, 1996).
Figure 2: Spectral Acceleration Map of 1995 Kobe, Japan Earthquake, Credit: U.S. Geologic Survey
Earthquake Impact
Occuring a year later on the exact same calendar day as the 1994 Northridge California Earthquake, the 1995 Kobe Earthquake completely decimated Kobe and the surrounding region. Causing extensive and catasrophic damage to the nations buildings and infrastructure, the earthquake claimed the lives of 5,426 people, injured 26,804 people, and rendered over 300,000 people homeless (Brebbia, 1996). Overall, 150,000 buildings collapsed or were irreparably damaged, and economic losses totaled nearly 20 trillion Yen (200 Billion US Dollars). With the financial and transportation infrastructure completely crippled, emergency response and recovery efforts were inefficient. One of the most damaging aspects of the earthquake was fire. In fact, more than 100 fires broke out within minutes after the earthquake (NCEER, 1995). In the end, 142 separate fires had been reported and approximately 1 million square meters of Kobe Japan had been burned (NCEER, 1995).
Infrastructure: Power, Telecommunications, Water, Waste Water, and Natural Gas Systems
Lifelines are always vulnerable to natural hazards such as earthquakes; however, their functionality must be sustained in order to guarantee public safety, economic vitality, and quality of life. Severe damage was sustained to the water, wastewater, and natural gas systems (Dickenson, 1996). Nearly all of Kobe was without services from water, waste water, and gas utilities. Ground failure resulted in severely damaged water pipelines as well as breaks in the underground gas distribution systems. Thirty five days passed before the water system was fully restored on the Port Island (Chung, 1996). As a result of the earthquake and subsuquent loss of power, water flow was reduced from the normal 6000 m^3/hr to just 1000 m^3/hr (Chung, 1996). Also, many of the water distribution reservoirs quickly emptied which seriously hindered fire suppression efforts. Pipelines were constructed with ductile iron material, which has a reputation for performing well during seismic events. Therefore, it is not surprising that the primary pipeline failures occurred at the joints where pipes seperated or pulled apart. The waste water system was badly damaged and services were temporarily curtailed. In fact, 3 of the 8 waste water plants in Kobe were damaged and 20 of 23 pump stations were damaged or at least inoperable due to power loss (Chung, 1996). In total, waste water system repairs amounted to an estimated 12 billion yen (Chung, 1996). For the most part, electric and telecommunication systems sustained functionality.
Hospitals
Hospitals provide critical care services, and are especially important during times of crisis and disaster. It is extremeley important to maintain operation of such facilities for the sake of current patients and incoming patients injured during disasters. Overall, Japan's hospital and critical care facilities performed well with a few exceptions. First, several hospitals reported loss of gas supply which prevented preparation of patient meals (Chung, 1996). Next, loss of water supply inhibited various operations. In fact, one hospital completely evacuated because the generators relied on the water supply (Chung, 1996). Also, in some cases the oxygen distribution systems were damaged. Finally, many hospitals reported equipment being moved or even falling.
Emergency Response / Preparedness
Overall, Japan displayed very little emergency response preparation. The scale and extent of destruction was not known for days, which slowed down recovery planning and efforts (NCEER, 1995). For days after the earthquake, there was little evidence of coordination amongst public safety and response crews (NCEER, 1995). The police, fire fighters, and medical teams were slow to reach hazardous areas. Perhaps the most shocking aspect of Japan's lack of preparedness was its neglect of building inspection. No official system existed for examining buildings or regulating entry into or around dangerous structures (NCEER, 1995). Finally, with the number of homeless people peaking at approximately 235,443, the shelters were overcrowded resulting in disease, influenza, and pneumonia.
Japan's Seismic Building Codes
History: Development of the Code
In order to fully understand how and why widespread structural failure occurred on January 17, 1995 in Kobe Japan, it is important to understand Japan's past and present approach to seismic design of buildings and civil engineering stuctures. In 1891, the Earthquake Investigation Committee was formed. Twenty four years later in 1915, a 0.1 seismic coefficient was introduced defining design seismic lateral design loads as the product of this (0.1) coefficient and the structures weight (Chung, 1996). In 1950, the Urban Building Law was replaced with the Building Standard Law and the seismic coefficient was increased to 0.2. In 1971 and 1981, two major updates in the seismic code were made. As a result, at the time of the Kobe Earthquake in 1995, there were three extremely different types of structures: buildings built prior to 1971, between 1971 and 1981, and buildings designed after 1981(Chung, 1996). In 1971, more strict requirements were established for transverse reinforcement size and spacing in reinforced concrete columns. In 1981, the seismic design coefficient was established as a variable as opposed to a constant (Chung, 1996). Provisions required calculation of the coefficient based on the structural vibration period. Another code change in 1981 was the emphasis on not just strength, but also ductility, life safety, and collapse mitigation.
Effect on Seismic Performance
As shown in Figure 3 below, there is positive correlation between the establishment of strict seismic design codes and improved seismic performance in Japan's structures. As will be thoroughly discussed in the Performance of Structures and Causes of Failure section, buildings constructed before 1971 and 1981 suffered the most severe damage.
Date of Construction
Percentage of Collapsed or Irreparably Damaged Buildings
(Extrapolated Values Based on Survey)
Before 1970
+55%
1971 to 1975
30%
1976 and 1980
10%
After 1980
None
Figure 3 (Chung, 2006, Page 87)
Performance of Structures and Causes of Failure
General Performance
Figure 4 (Chung, 2006)
Traditional Wood Frame Houses
Historically, wood frame structures have performed well during seismic events throughout the world. Unfortunately, several common design decisions and substandard material and construction methods resulted in widespread structural failure throughout Kobe Japan. More than 10% of the wood frame houses totally or partially collapsed. Houses built it the 50s and 60s accounted for the majority of collapsed homes. Two poular types of wood frame stuctures exist in Japan which are Shinkabe houses and Ohkabe houses (Chung, 1996).
Shinkabe houses are wood frame structures with exterior facade systems comprised of stucco and virtually no structural resistance. They feature post and beam framing and mudfilled bamboo walls are relied on for lateral resistance. The connections are primarily wood joiners instead of nails and connectors. The defining characteristics of Ohkabe houses are columns joined by wooden slats and reinforced with light wooden cross braces. Overall, wood frame houses performed poorly and account for many of the collapsed buildings and lives lost.
The problems with the wood frame houses were fundamental. No plywood or particleboard shearwalls were used, and heavy ceramic tile roofs were used. During earthquakes, the magnitude of lateral forces is positively correlated to the mass of the structure. Heavy tile roofs introduce large masses at a higher building elevation resulting in large overturning moments and greater shear demands on the supporting wood framing. For the most part, in one and two story wooden post and beam construction, the lack of bracing and heavy roofs forced the first story to collapse. Overall, approximately 55,000 houses collapsed and 35,000 were severely damaged (Chung, 1996). Shinkabe construction is more modern and performed better than Ohkabe structures. A third type of wood structure is the more recent prefabricated timber house, which performed very well attesting to the power of the updated building codes. With more stringent code requirements, timber homes built after 1981 performed relatively well as did the few American style studwall homes. Many residents choose to construct roofs out of heavy tile set in mud to withstand hurricane winds. However lighter synthetic roofs are the best option for seismic performance since the structures weight dictates the magnitude of seismic loads.
Reinforced Concrete Buildings
Although the post 1981 Reinforced Concrete (RC) buildings performed very well, older RC buildings did not. Well over 100 Midrise concrete buildings constructed in the 1960s and 1970s totally collapsed. The 1981 code improvements to focus on life safety and collapse mitigation made a big difference. For example, the entire sixth floor collapsed in the 1960 Kobe City Hall; however, the the adjacent new city hall is a 16 story RC building and it was barely damaged.
The collapse of RC buildings can be attributed to a wide variety of failure mechanisms. For building built prior to 1971, middle story collapse was predominant due to brittle shear failure of RC columns. The underlying cause of middle story collapse in older RC buildings was poor detailing and low ductility. Other failure patterns identified in RC buildings was lack of column transveres reinforcement, smooth reinforcement, failure of longitudinal reinforcement splices made by gas pressure welding. In many cases, the soft story design was the leading cause of collapse. Soft story design is defined by a weak, flexible, and open first story with too few columns. The upper floors are stiffer and rotate on the soft first story resulting in column failure and ultiamtely collapse. Down below, Figure 8 shows a multi story RC building that suffered total collapse at an intermediate level soft story.
In many cases RC buildings failed becuase the combined effects of soft story design and poor column detailing. Especially in older structures, no column hoop ties or excessive column tie spacing caused loss of cover, shattering of column core, and bucking of longitdunial column reinfrocement. With proper detailing, columns cores may completely detriorate under excessive seismic racking; however, the column reinforcement maintains its structural integrity and prevents collapse.
Another common failure mechanism was vertical and plan irregulariteis in regards to mass and stiffness distribution. Many multi-level RC strucures collapsed at upper levels due to stiffness irregularites (Chung, 1996). For example a 9 story box shaped structure had shear walls on 3 sides and moment frames on the fourth side, and this irregularity in stiffness resulted in collapse. During the earthquake, torsional motion was induced leaning the building towards the moment frames which resulted in shear failure of the moment frame columns and tension failure in the shear wall connections (Chung, 1996).
One of the more universal failure mechanisms identified in collapsed RC buildings was the strong girder-weak column system which is shown in Figure 5. Also, in many cases excessive openings in walls resulted in reduced cross sectional areas which caused major diagonal tension cracks during the earthquake. It is important to note that RC frame construction performed much worse than RC bearing wall and wall girder structures.
Figure 5: Strong Beam -Weak Column. Contribution of National Institute of Standards and Technology (Chung, 1996)
Steel Reinforced Concrete
Much like the RC buildings, Steel Reinforced Concrete (SRC) structures peformed better when built after 1981. In cases where dual shear wall and SRC moment frames were employed, large diagonal tension cracks occured in the shear walls. When correctly designed, SRC buildings have increased ductility. However, in the past it was common practice to discontinue embedded steel members above higher levels. Consequently, an uneven stiffness and strength distribution is established. Also, many of the SRC buildings before 1981 were designed for uniform lateral forces instead of the more realistic modal response of structures. As a result, upper level stories were weaker and undersized for both shear and flexural resistance. Entire failure of upper level stories was typical. The pre 1917 open web SRC design was way outpeformed by the more reliable Full web SRC design.
Steel Frame Buildings
The same trend of newer buildings outperforming older buildings was again detected. Extensive damage to steel structures was reported; however, relatively few collapsed and paths of egress remained functional. Failure of welds was by far the most common form of damage, and was reminiscent of the connection failures observed during the Northridge Earthquake in 1994 (Chung, 1996). Older steel buildings performed badly with round bars for beams and built up columns made with channel sections.
Column-beam sections accounted for the majority of steel frame building failures. Especiallly when built before 1981, Japan's steel frame buildings and no stiffening plates and expereienced fracture at fillet welds . Other modes of failure included fracture of column spices, bolt fracture at beam splices, and damaging interstory drifts resulting from damaged bracing members. Finally, column base damage and pull out of anchor bolts was common.
Highways and Bridges
The damage to elevated expressways and bridges was widespread and catasrophic. Typical damage included shear and flexural failure of concrete columns, buckling of steel columns, foundation movement from ground failure, and unseated girders. Both new and old bridges performed poorly creating the need for code revisions and retrofitting efforts. Many of the bridges were designed to resist seismic loads through brute strength, whereas the US relies on strength and ductility and fllexibility (Blakeslee, 1995).
The fundamental flaw in bridge and highway construction was emphasis on brute strength. Massive round RC pillars were ineffectively relied upon for seismic resistance. Perhaps the most dramatic failure was the complete destruction of the Hanshin Expressway. A 550 yard span completely overturned, which is pictured in Figure 6 (Pollack, 1995). The elevated highway overturned starting at a section where there is a transition from steel griders to RC girders (Pollack, 1995). The additional weight from the concrete likely caused the collapse. The Hanshin expressway failed in two other locations as well. In the first location, the columns moved in displaced in opposite directions dislodging the roadbed from supports. In the other location, there was shear and flexure failure in the hammerhead RC piers. It is interesting to note that the Hanshin Expressway was supported by a single row of columns, which allowed the roadbed to overturn (Pollack, 1995). As in the collapse of the Shinkansen bullet train, many of the elevated viadcuts built prior to 1971 collapsed when the columns failed in shear.
Figure 6: Permission Pending (Kitigawa, 2004)
Daikai Subway Station
The collapse of the Daikai Subway Station was of great significance, for it represented the first major failure of a modern cut and cover subway station (Parra-Montesinos Gustavo, 2006). The structure is an underground RC box structure. While under shear and vertical loading, over 30 central RC columns collapsed because they did not have sufficient confining steel (Narra-Montesinos Gustavo, 2006). The two basic types of transverse reinforcement used were 90 degree perimeter hooks and zig zag stirrups supporting every other longitudinal bar (Parra-Montesinos, 2006). With large spacings, the transverse reinforcement resulted in reduced drift capacity under the high axial loads from overburden soils. Additionally, the transverse reinforcement was not sufficient enough to withstand shear reversals or give proper lateral support to the longitudinal reinforcement (Parra-Montesinos Gustavo, 2006).
Geotechnical Factors
In addition to all structural engineering design flaws and substandard construction methods, geological condtions played a major role in dictating the distribution and severity of structural damage. Large sections of Kobe Port Island and Rokko Island are manmade landfills (Dickenson,1996). The entire soil profile of Kobe City consists of deposited sands at the surface, then a 10-15 meter soft marine clay layer, followed by a thick layer of gravel and sand, which is underlaid by a stiff Pleistocene Clay (Brebbia, 1996). Figure 7 shows the original soil profile, which now has a loose layer of deposited sands on the top. These artificial islands were created by placing sandy decomposed granite over top of compressible marine clay. As a result, many of Kobe Japan's structures are constructed on top of loose soft land which presents two major seismic hazards.
First, an unprceedented amount of liquefaction occurred. During the earthquake, violent ground shaking induced the liquefaction of an estimated 17 square kilometers of loose soils (Chung, 1996). As the soft soils expelled large quantitites of sand and water, the liquefaction process amplified the peak ground acceleration to nearly .8g in some locations. Second, setttlement issues associated with the soft soils contributed to poor seismic performance of buildings. With high moisture content in the marine clays, the deposited sand and granite generated an average settlement of 4 meters and settlement is still occurring (Brebbia, 1996).
For the most part, structures located along rivers or in coastal regions were the most affected by these geological factors. Ground failure beneath many of the port warehouses displaced foundation piles breaking the pile caps. Liquefaction was responsible for an estimated 50 linear kilometers of retaining wall damage (NCEER, 1995). Both islands experienced liquefaction and sesimic induced ground settlement of 30-60 cm. Extensive shoreline ground failure was reported along the west coast of the Port Island and along the Southern section of Rokko Island. Granular fill behind caissons of shoreline structures settled which tilted and damaged the caissons (Brebbia, 1996). Satellite imagery revealed the Port Island ground failure included 2-3 meters of horizontal movement as well as 1 meter of settlement (Brebbia, 1996). Heavy damages to port structures and harbor facilities was sustained due to these ground failures (Dickenson, 1996).
Figure 7: General Soil Profile. Suseptible to Liquefaction (Brebbia, 1996)
Conclusion
The powerful 1995 Kobe Japan Earthquake was measured 7.2 (JMA) in Magnitude,and struck in a geologic region that is seismically hazardous. Even with liquefaction amplifying peak ground accelerations, the widespread destruction and mass loss of life was avoidable. Most of the collapsed buildings were constructed prior to 1971 and 1981 when major seismic code revisions were made. Therefore it is clear Japan is headed in the right direction with the development of their seismic codes; however, faillure to retrofit substandard structural systems is evident and likely to remain a concerny in any future earthquakes of similar magnitude. The predominant modes of failure included intermediate story level collapse, soft story first level collapse, total collapse, shear failure in poorly detailed RC columns, and fracture of fillet welds. Key design flaws included excessive column ties spacing, smooth reinforcement, top heavy structures, weak or soft story design, and irregular distribution of stiffness and strength in both the vertical and horizontal directions.
Figure 8: Intermediate Level Soft Story Collapse. Contribution of the National Institute of Standards and Technology (Chung, 1996)
Bibliography
ReportAkai, Koichi and Jonathan D. Bray. (1995) "Geotechnical Reconnaisance of The Effects of The January 17, 1995, Hyogoken Nanbu Earthquake, Japan." Earthquake Engineering Research Center Report. July, 1995http://www.ce.berkeley.edu/geo/research/Kobe/KobeReport/contents.html
Generated by a team comprised of University of California at Berkley faculty members and the Earthquake Engineering Research Center, this report provides extensive text and images detailing the seismic performance of port facilities, levees, dams, bridges, highways, and underground rapid transit systems during the 1995 Kobe Earthquake. A large portion of the report focuses on soil and foundation conditions, as well as the role of liquefaction as a failure mechanism.
Journal
Azizinamini, Atorod and S. K. Ghosh. (1997) "Steel Reinforced Concrete Structures in 1995 Hyogoken Nanbu Earthquake." Journal of Structural Engineering, Vol. 123, no. 8 (1997): 986-992.
Written by University of Nebraska Civil Engineering Professor Atrod Azizinamini, this journal article discusses the observations of the damage sustained by steel reinforced concrete structures during the Kobe Earthquake. Also, the article explores possible failure mechanisms including excessive column tie spacing, inadequate base plate detailing, improper hooks, and vertical vertical discontinuity of building stiffnesses.
A New York Times article written by Sandra Blakeslee just one week after the Kobe Earthquake took place. This article discusses initial reactions and evaluations of the investigating engineers. The article discusses differences in US and Japan structural design. Specifically, the article discusses why newer buildings faired better than oloder buildings, and identifies several significant design flaws present throught Japan's Column designs.
This New York Times article breifly explains how Japanese structural engineering practices differ from the US. The importance of balance between flexibility and strength are discussed, and a summary of bridges and building performances is provided.
Book
Brebbia, C. A. (1996) The Kobe Earthquake: Geodynamical Aspects. Southampton, UK: Computational Mechanics Publications, 1996.
The 1st volume in the Advances in Earthquake Engineering Series, this book primarily focuses on the soil and seismotectonic characteristics of the Kobe region. Soil amplification, ground motion, landslides, liquefaction, and other ground failures are identified and discussed as major contributors to the severe damage resulting from the Kobe Earthquake. A chapter is also dedicated to explaining the structural damage sustained by buildings and the progress made in regards to the restoration process.
Book/WebsiteChung, Riley M. (1996) The January 17, 1995, Hyogoken Nanbu (Kobe) Earthquake: Performance of Structures, Lifelines, and Fire Protection Systems. Gaithersbrug, MD: The Institute, 1996http://www.fire.nist.gov/bfrlpubs/build96/PDF/b96002.pdf
A National Institute of Standards and Technology Special Report which summarizes the key findings and information gathered by an 18 man reconaissance team that explored the Kobe Japan Aftermath for 7 days. The report covers fires, seismology, geology issues, as well as background regarding the loss of life and economic ramifications. However, the primary focus of the report pertains to identifying and explaining the types of failure mechanisms most commonly observed for each type of building. Wood frame construction, RC bearing wall, RC frame construction, SRC, and steel construction are all investigated and supplemented with images of failed structures.
BookDickenson, S.E. and Stuart D. Weaver. (1996) Hyogoken Nanbu Earthquake of January 17, 1995. New York, NY: American Society of Civil Engineers, 1996.
This book provides a detailed description of the seismic performance of various ports throughout the Kobe region including the Osaka Port and the Kasai Airport. In addition, the performance of parking structures, highways, metal, wood, and brittle concrete frame buildings are examined and explained.
JournalKitagawa, Yoshikazu and Hisahiro Hiraishi. (2004) "Overview of the 1995 Hyogoken Nanbu Earthquake and Proposals for Earthquake Mitigation Measures." Journal of Japan Association for Earthquake Engineering, Vol. 4, no. 3 (2004)http://www.jaee.gr.jp/stack/submit-j/v04n03/data/pdf/1.pdf
Written by Yoshikazu Kitagawa of JAEE, this journal article outlines the overall destruction resulting from the Kobe Earthquake. Topics explored include geological conditions, geotechnical apsects, the role of strong ground motions, and the structural damage suffered by buildings, infrastructure, and lifelines.
Journal
Parra-Montesinos, Gustavo J. and Antonio Bobet. (2006) "Evaluation of Soil Structure Intereaction and Structural Collapse in Daikai Subway Station During Kobe Earthquake." American Concrete Institute Structural Journal, Vol. 103, no. 1: 113.
From the American Concrete Institute Structural Journal, this article highlights the prominent role of soil-structure interaction in causing the Daikai Subway Station to collapse during the 1995 Kobe Earthquake. The article explains the discrepancy between design load and drift with respect to actual drift demands imposed on the structure. The article also evaluates the seismic behavior of various other underground structures, and discusses major design flaws in RC columns which collapsed via shear failure all across the Kobe region. Statistics regarding the amount and severity of failed columns througout the underground transit systems is also included.
Written by Andrew Pollack, this New York Times article provides a qualitative description of structural failures in department stores, high rises, highways, bridges, and offfice buildings. The article disscusses how severity of damage was distributed sporadically across the region, and discusses why this signifies improper seismic designs. Also, the article discusses the history of Japan's building code and their design approach with regards to acceleration as a percentage of gravity versus tolerable/expected level of damage.
State University of New York at Buffalo. (1995) "Report: A Summary of Earthquake Reconaissance Efforts of The National Center for Earthquake Engineering Research Center."
Composed by the National Center for Earthquake Engineering Reserach Center, this electronic report examines performance of steel, wood, and concrete buildings as well as roads, highways, ports, and harbors. Liquefaction induced lateral displacement is examined as a culpirt for the majority of foundation system failures. A section also discusses the effectiveness of Japan's Building Code based upon the correlation between structural damage and the governing building code for each structure. Finally, social and economic impacts of the Kobe Earthquake are discussed.
1995 Kobe, Japan Earthquake
Key Terms: Earthquake, Seismology, Liquefaction, Soft Story Effects, RC Column Shear Failures
Table of Contents
| Introduction | Seismology | Earthquake Impact | Japan's Seismic Building Codes | Performance of Structures and Causes of Failure | Geotechnical Factors | Conclusion | Bibliography
Introduction
On January 17, 1995 at approximately 5:47 am local time, a 7.2 Magnitude (Japan Meteorlogical Agency Scale) Earthquake decimated Kobe, Japan and the surrounding region. An entire order of magnitude stronger than the Northridge Earthquake, the Kobe Earthquake (Hyogo Ken Nanbu Earthquake) lasted for nearly 15 seconds claiming the lives of over 5,000 people, injuring more than 26,000 others, displacing 300,000 residents, destroying over 150,000 buildings, and accumulating an estimated 20 trillion Yen (200 Billion US Dollars) in economic losses (Akai, 1995). The epicenter was located at the northern part of Awaji Island, just 20 km from the densely developed urban city of Kobe, which is home to nearly 1.5 million people (Akai, 1995). With extensive structural damage to ports, harbors, dams, levees, underground transit systems, highways, bridges, and buildings, the aftermath of the Kobe Japan Earthquake was a living laboratory for engineers to explore and learn from. The structural damage depicted in Figure 1 was widespread throughout Japan. Although a wide array of failure mechanisms were identified throughout the region, the primary design flaws include the following:
Seismology
Kobe Japan is located in an earthquake prone region dominated by complex faulting systems (Akai, 1995). The Hanshin region and Awaji Island are affected by the subducting Philippine Sea Plate and the Plastic Plate. The 1995 Kobe Japan earthquake was initiated by the release of accumulated stress generated by the subsiding, uplifting, and tilting oftectonic plates. Specifically, the earthquake source can be traced to three tectonic subevents that occurred along the Nojima Fault, which is directly beneath the Akashi Strait. These three events included two strike slip ruptures and a bilateral rupture. Japan Meteorological Agency reported a registered magnitude of 7.2. The epicenter was located at the northern part of Awaji island, just 20 km from the densely urbanized city of Kobe, Japan (Kitigawa, 2004). As shown in Figure 2, the northern coast of Japan experienced the greatest ground shaking. With a 16 km focal depth and a 15 second duration, the earthquake caused 1.7 m lateral surface displacement and 1.0 m vertical ground displacement. Peak ground shaking was .5g, but a limited number of locations experienced peak ground motion that exceeded .8g (Chung, 1996). The overall rupture length was approximately 30 to 50 kilometers (Dickenson, 1996).
Earthquake Impact
Occuring a year later on the exact same calendar day as the 1994 Northridge California Earthquake, the 1995 Kobe Earthquake completely decimated Kobe and the surrounding region. Causing extensive and catasrophic damage to the nations buildings and infrastructure, the earthquake claimed the lives of 5,426 people, injured 26,804 people, and rendered over 300,000 people homeless (Brebbia, 1996). Overall, 150,000 buildings collapsed or were irreparably damaged, and economic losses totaled nearly 20 trillion Yen (200 Billion US Dollars). With the financial and transportation infrastructure completely crippled, emergency response and recovery efforts were inefficient. One of the most damaging aspects of the earthquake was fire. In fact, more than 100 fires broke out within minutes after the earthquake (NCEER, 1995). In the end, 142 separate fires had been reported and approximately 1 million square meters of Kobe Japan had been burned (NCEER, 1995).Infrastructure: Power, Telecommunications, Water, Waste Water, and Natural Gas Systems
Lifelines are always vulnerable to natural hazards such as earthquakes; however, their functionality must be sustained in order to guarantee public safety, economic vitality, and quality of life. Severe damage was sustained to the water, wastewater, and natural gas systems (Dickenson, 1996). Nearly all of Kobe was without services from water, waste water, and gas utilities. Ground failure resulted in severely damaged water pipelines as well as breaks in the underground gas distribution systems. Thirty five days passed before the water system was fully restored on the Port Island (Chung, 1996). As a result of the earthquake and subsuquent loss of power, water flow was reduced from the normal 6000 m^3/hr to just 1000 m^3/hr (Chung, 1996). Also, many of the water distribution reservoirs quickly emptied which seriously hindered fire suppression efforts. Pipelines were constructed with ductile iron material, which has a reputation for performing well during seismic events. Therefore, it is not surprising that the primary pipeline failures occurred at the joints where pipes seperated or pulled apart. The waste water system was badly damaged and services were temporarily curtailed. In fact, 3 of the 8 waste water plants in Kobe were damaged and 20 of 23 pump stations were damaged or at least inoperable due to power loss (Chung, 1996). In total, waste water system repairs amounted to an estimated 12 billion yen (Chung, 1996). For the most part, electric and telecommunication systems sustained functionality.
Hospitals
Hospitals provide critical care services, and are especially important during times of crisis and disaster. It is extremeley important to maintain operation of such facilities for the sake of current patients and incoming patients injured during disasters. Overall, Japan's hospital and critical care facilities performed well with a few exceptions. First, several hospitals reported loss of gas supply which prevented preparation of patient meals (Chung, 1996). Next, loss of water supply inhibited various operations. In fact, one hospital completely evacuated because the generators relied on the water supply (Chung, 1996). Also, in some cases the oxygen distribution systems were damaged. Finally, many hospitals reported equipment being moved or even falling.
Emergency Response / Preparedness
Overall, Japan displayed very little emergency response preparation. The scale and extent of destruction was not known for days, which slowed down recovery planning and efforts (NCEER, 1995). For days after the earthquake, there was little evidence of coordination amongst public safety and response crews (NCEER, 1995). The police, fire fighters, and medical teams were slow to reach hazardous areas. Perhaps the most shocking aspect of Japan's lack of preparedness was its neglect of building inspection. No official system existed for examining buildings or regulating entry into or around dangerous structures (NCEER, 1995). Finally, with the number of homeless people peaking at approximately 235,443, the shelters were overcrowded resulting in disease, influenza, and pneumonia.
Japan's Seismic Building Codes
History: Development of the CodeIn order to fully understand how and why widespread structural failure occurred on January 17, 1995 in Kobe Japan, it is important to understand Japan's past and present approach to seismic design of buildings and civil engineering stuctures. In 1891, the Earthquake Investigation Committee was formed. Twenty four years later in 1915, a 0.1 seismic coefficient was introduced defining design seismic lateral design loads as the product of this (0.1) coefficient and the structures weight (Chung, 1996). In 1950, the Urban Building Law was replaced with the Building Standard Law and the seismic coefficient was increased to 0.2. In 1971 and 1981, two major updates in the seismic code were made. As a result, at the time of the Kobe Earthquake in 1995, there were three extremely different types of structures: buildings built prior to 1971, between 1971 and 1981, and buildings designed after 1981(Chung, 1996). In 1971, more strict requirements were established for transverse reinforcement size and spacing in reinforced concrete columns. In 1981, the seismic design coefficient was established as a variable as opposed to a constant (Chung, 1996). Provisions required calculation of the coefficient based on the structural vibration period. Another code change in 1981 was the emphasis on not just strength, but also ductility, life safety, and collapse mitigation.
Effect on Seismic Performance
As shown in Figure 3 below, there is positive correlation between the establishment of strict seismic design codes and improved seismic performance in Japan's structures. As will be thoroughly discussed in the Performance of Structures and Causes of Failure section, buildings constructed before 1971 and 1981 suffered the most severe damage.
(Extrapolated Values Based on Survey)
Performance of Structures and Causes of Failure
General PerformanceTraditional Wood Frame Houses
Historically, wood frame structures have performed well during seismic events throughout the world. Unfortunately, several common design decisions and substandard material and construction methods resulted in widespread structural failure throughout Kobe Japan. More than 10% of the wood frame houses totally or partially collapsed. Houses built it the 50s and 60s accounted for the majority of collapsed homes. Two poular types of wood frame stuctures exist in Japan which are Shinkabe houses and Ohkabe houses (Chung, 1996).
Shinkabe houses are wood frame structures with exterior facade systems comprised of stucco and virtually no structural resistance. They feature post and beam framing and mudfilled bamboo walls are relied on for lateral resistance. The connections are primarily wood joiners instead of nails and connectors. The defining characteristics of Ohkabe houses are columns joined by wooden slats and reinforced with light wooden cross braces. Overall, wood frame houses performed poorly and account for many of the collapsed buildings and lives lost.
The problems with the wood frame houses were fundamental. No plywood or particleboard shearwalls were used, and heavy ceramic tile roofs were used. During earthquakes, the magnitude of lateral forces is positively correlated to the mass of the structure. Heavy tile roofs introduce large masses at a higher building elevation resulting in large overturning moments and greater shear demands on the supporting wood framing. For the most part, in one and two story wooden post and beam construction, the lack of bracing and heavy roofs forced the first story to collapse. Overall, approximately 55,000 houses collapsed and 35,000 were severely damaged (Chung, 1996). Shinkabe construction is more modern and performed better than Ohkabe structures. A third type of wood structure is the more recent prefabricated timber house, which performed very well attesting to the power of the updated building codes. With more stringent code requirements, timber homes built after 1981 performed relatively well as did the few American style studwall homes. Many residents choose to construct roofs out of heavy tile set in mud to withstand hurricane winds. However lighter synthetic roofs are the best option for seismic performance since the structures weight dictates the magnitude of seismic loads.
Although the post 1981 Reinforced Concrete (RC) buildings performed very well, older RC buildings did not. Well over 100 Midrise concrete buildings constructed in the 1960s and 1970s totally collapsed. The 1981 code improvements to focus on life safety and collapse mitigation made a big difference. For example, the entire sixth floor collapsed in the 1960 Kobe City Hall; however, the the adjacent new city hall is a 16 story RC building and it was barely damaged.
The collapse of RC buildings can be attributed to a wide variety of failure mechanisms. For building built prior to 1971, middle story collapse was predominant due to brittle shear failure of RC columns. The underlying cause of middle story collapse in older RC buildings was poor detailing and low ductility. Other failure patterns identified in RC buildings was lack of column transveres reinforcement, smooth reinforcement, failure of longitudinal reinforcement splices made by gas pressure welding. In many cases, the soft story design was the leading cause of collapse. Soft story design is defined by a weak, flexible, and open first story with too few columns. The upper floors are stiffer and rotate on the soft first story resulting in column failure and ultiamtely collapse. Down below, Figure 8 shows a multi story RC building that suffered total collapse at an intermediate level soft story.
In many cases RC buildings failed becuase the combined effects of soft story design and poor column detailing. Especially in older structures, no column hoop ties or excessive column tie spacing caused loss of cover, shattering of column core, and bucking of longitdunial column reinfrocement. With proper detailing, columns cores may completely detriorate under excessive seismic racking; however, the column reinforcement maintains its structural integrity and prevents collapse.
Another common failure mechanism was vertical and plan irregulariteis in regards to mass and stiffness distribution. Many multi-level RC strucures collapsed at upper levels due to stiffness irregularites (Chung, 1996). For example a 9 story box shaped structure had shear walls on 3 sides and moment frames on the fourth side, and this irregularity in stiffness resulted in collapse. During the earthquake, torsional motion was induced leaning the building towards the moment frames which resulted in shear failure of the moment frame columns and tension failure in the shear wall connections (Chung, 1996).
One of the more universal failure mechanisms identified in collapsed RC buildings was the strong girder-weak column system which is shown in Figure 5. Also, in many cases excessive openings in walls resulted in reduced cross sectional areas which caused major diagonal tension cracks during the earthquake. It is important to note that RC frame construction performed much worse than RC bearing wall and wall girder structures.
Steel Reinforced Concrete
Much like the RC buildings, Steel Reinforced Concrete (SRC) structures peformed better when built after 1981. In cases where dual shear wall and SRC moment frames were employed, large diagonal tension cracks occured in the shear walls. When correctly designed, SRC buildings have increased ductility. However, in the past it was common practice to discontinue embedded steel members above higher levels. Consequently, an uneven stiffness and strength distribution is established. Also, many of the SRC buildings before 1981 were designed for uniform lateral forces instead of the more realistic modal response of structures. As a result, upper level stories were weaker and undersized for both shear and flexural resistance. Entire failure of upper level stories was typical. The pre 1917 open web SRC design was way outpeformed by the more reliable Full web SRC design.
The same trend of newer buildings outperforming older buildings was again detected. Extensive damage to steel structures was reported; however, relatively few collapsed and paths of egress remained functional. Failure of welds was by far the most common form of damage, and was reminiscent of the connection failures observed during the Northridge Earthquake in 1994 (Chung, 1996). Older steel buildings performed badly with round bars for beams and built up columns made with channel sections.
Column-beam sections accounted for the majority of steel frame building failures. Especiallly when built before 1981, Japan's steel frame buildings and no stiffening plates and expereienced fracture at fillet welds . Other modes of failure included fracture of column spices, bolt fracture at beam splices, and damaging interstory drifts resulting from damaged bracing members. Finally, column base damage and pull out of anchor bolts was common.
Highways and Bridges
The damage to elevated expressways and bridges was widespread and catasrophic. Typical damage included shear and flexural failure of concrete columns, buckling of steel columns, foundation movement from ground failure, and unseated girders. Both new and old bridges performed poorly creating the need for code revisions and retrofitting efforts. Many of the bridges were designed to resist seismic loads through brute strength, whereas the US relies on strength and ductility and fllexibility (Blakeslee, 1995).
The fundamental flaw in bridge and highway construction was emphasis on brute strength. Massive round RC pillars were ineffectively relied upon for seismic resistance. Perhaps the most dramatic failure was the complete destruction of the Hanshin Expressway. A 550 yard span completely overturned, which is pictured in Figure 6 (Pollack, 1995). The elevated highway overturned starting at a section where there is a transition from steel griders to RC girders (Pollack, 1995). The additional weight from the concrete likely caused the collapse. The Hanshin expressway failed in two other locations as well. In the first location, the columns moved in displaced in opposite directions dislodging the roadbed from supports. In the other location, there was shear and flexure failure in the hammerhead RC piers. It is interesting to note that the Hanshin Expressway was supported by a single row of columns, which allowed the roadbed to overturn (Pollack, 1995). As in the collapse of the Shinkansen bullet train, many of the elevated viadcuts built prior to 1971 collapsed when the columns failed in shear.
Daikai Subway Station
The collapse of the Daikai Subway Station was of great significance, for it represented the first major failure of a modern cut and cover subway station (Parra-Montesinos Gustavo, 2006). The structure is an underground RC box structure. While under shear and vertical loading, over 30 central RC columns collapsed because they did not have sufficient confining steel (Narra-Montesinos Gustavo, 2006). The two basic types of transverse reinforcement used were 90 degree perimeter hooks and zig zag stirrups supporting every other longitudinal bar (Parra-Montesinos, 2006). With large spacings, the transverse reinforcement resulted in reduced drift capacity under the high axial loads from overburden soils. Additionally, the transverse reinforcement was not sufficient enough to withstand shear reversals or give proper lateral support to the longitudinal reinforcement (Parra-Montesinos Gustavo, 2006).
Geotechnical Factors
First, an unprceedented amount of liquefaction occurred. During the earthquake, violent ground shaking induced the liquefaction of an estimated 17 square kilometers of loose soils (Chung, 1996). As the soft soils expelled large quantitites of sand and water, the liquefaction process amplified the peak ground acceleration to nearly .8g in some locations. Second, setttlement issues associated with the soft soils contributed to poor seismic performance of buildings. With high moisture content in the marine clays, the deposited sand and granite generated an average settlement of 4 meters and settlement is still occurring (Brebbia, 1996).
For the most part, structures located along rivers or in coastal regions were the most affected by these geological factors. Ground failure beneath many of the port warehouses displaced foundation piles breaking the pile caps. Liquefaction was responsible for an estimated 50 linear kilometers of retaining wall damage (NCEER, 1995). Both islands experienced liquefaction and sesimic induced ground settlement of 30-60 cm. Extensive shoreline ground failure was reported along the west coast of the Port Island and along the Southern section of Rokko Island. Granular fill behind caissons of shoreline structures settled which tilted and damaged the caissons (Brebbia, 1996). Satellite imagery revealed the Port Island ground failure included 2-3 meters of horizontal movement as well as 1 meter of settlement (Brebbia, 1996). Heavy damages to port structures and harbor facilities was sustained due to these ground failures (Dickenson, 1996).
Conclusion
Bibliography
ReportAkai, Koichi and Jonathan D. Bray. (1995) "Geotechnical Reconnaisance of The Effects of The January 17, 1995, Hyogoken Nanbu Earthquake, Japan." Earthquake Engineering Research Center Report. July, 1995http://www.ce.berkeley.edu/geo/research/Kobe/KobeReport/contents.html
Journal
Azizinamini, Atorod and S. K. Ghosh. (1997) "Steel Reinforced Concrete Structures in 1995 Hyogoken Nanbu Earthquake." Journal of Structural Engineering, Vol. 123, no. 8 (1997): 986-992.
http://cedb.asce.org/cgi/WWWdisplay.cgi?107504Written by University of Nebraska Civil Engineering Professor Atrod Azizinamini, this journal article discusses the observations of the damage sustained by steel reinforced concrete structures during the Kobe Earthquake. Also, the article explores possible failure mechanisms including excessive column tie spacing, inadequate base plate detailing, improper hooks, and vertical vertical discontinuity of building stiffnesses.
NewspaperBlakeslee, Sandra. (1995) "Brute Strength Couldn't Save Kobe Structures." New York Times, January 25, 1995http://www.nytimes.com/1995/01/25/world/brute-strength-couldn-t-save-kobe-structures.html?scp=2&sq=Kobe+Japan+Earthquake+structural+engineering&st=nyt
NewspaperBlakeslee, Sandra. (1995) "QUAKE IN JAPAN: Japanese and U.S. Ways of Structural Engineering Differ." New York Times, January 18, 1995.http://select.nytimes.com/gst/abstract.html?res=F60617F93F5D0C7B8DDDA80894DD494D81&scp=32&sq=Kobe%20Japan%20Earthquake%20building%20collapses&st=cse
Book
Brebbia, C. A. (1996) The Kobe Earthquake: Geodynamical Aspects. Southampton, UK: Computational Mechanics Publications, 1996.
The 1st volume in the Advances in Earthquake Engineering Series, this book primarily focuses on the soil and seismotectonic characteristics of the Kobe region. Soil amplification, ground motion, landslides, liquefaction, and other ground failures are identified and discussed as major contributors to the severe damage resulting from the Kobe Earthquake. A chapter is also dedicated to explaining the structural damage sustained by buildings and the progress made in regards to the restoration process.
Book/WebsiteChung, Riley M. (1996) The January 17, 1995, Hyogoken Nanbu (Kobe) Earthquake: Performance of Structures, Lifelines, and Fire Protection Systems. Gaithersbrug, MD: The Institute, 1996http://www.fire.nist.gov/bfrlpubs/build96/PDF/b96002.pdf
BookDickenson, S.E. and Stuart D. Weaver. (1996) Hyogoken Nanbu Earthquake of January 17, 1995. New York, NY: American Society of Civil Engineers, 1996.
- This book provides a detailed description of the seismic performance of various ports throughout the Kobe region including the Osaka Port and the Kasai Airport. In addition, the performance of parking structures, highways, metal, wood, and brittle concrete frame buildings are examined and explained.
http://books.google.com/books?id=bMGaNQUva2EC&printsec=frontcover&dq=Dickinson,+S.E.+and+stuart+D.+Werner.+(1996)+Hyogo+Ken+Earthquake+of+January&source=bl&ots=ef1lon-CD4&sig=DEKO2KPATmBPASqaftLC0IL02Nw&hl=en&ei=zakLTaXJN4P-8Aaps6z6DQ&sa=X&oi=book_result&resnum=1&ved=0CBMQ6AEwAA#v=onepage&q&f=trueJournalKitagawa, Yoshikazu and Hisahiro Hiraishi. (2004) "Overview of the 1995 Hyogoken Nanbu Earthquake and Proposals for Earthquake Mitigation Measures." Journal of Japan Association for Earthquake Engineering, Vol. 4, no. 3 (2004)http://www.jaee.gr.jp/stack/submit-j/v04n03/data/pdf/1.pdf
Journal
Parra-Montesinos, Gustavo J. and Antonio Bobet. (2006) "Evaluation of Soil Structure Intereaction and Structural Collapse in Daikai Subway Station During Kobe Earthquake." American Concrete Institute Structural Journal, Vol. 103, no. 1: 113.
http://findarticles.com/p/articles/mi_qa5310/is_200601/ai_n21408247/?tag=content;col1From the American Concrete Institute Structural Journal, this article highlights the prominent role of soil-structure interaction in causing the Daikai Subway Station to collapse during the 1995 Kobe Earthquake. The article explains the discrepancy between design load and drift with respect to actual drift demands imposed on the structure. The article also evaluates the seismic behavior of various other underground structures, and discusses major design flaws in RC columns which collapsed via shear failure all across the Kobe region. Statistics regarding the amount and severity of failed columns througout the underground transit systems is also included.
NewspaperPollack, Andrew. (1995) "QUAKE IN JAPAN: THE ENGINEERING; Damage Survey Points to Design As a Weakness." New York Times, January 19, 1995.http://www.nytimes.com/1995/01/19/world/quake-in-japan-the-engineering-damage-survey-points-to-design-as-a-weakness.html?scp=1&sq=Earthquake%20Engineering%20Kobe%20Japan&st=cse&pagewanted=2
USGS Report / Website
Silva, Walter and Bob Darragh (2001). "Incorporation of Earthquake Source, Propogation Path, and Site Uncertainties into Assessment of Liquefaction Potential; Phase 1, Validation." United States Geologic Survey.http://search.usgs.gov/results.html?cx=005083607223377578371%3Ab5ixbbpqpx0&cof=FORID%3A11&q=incorporation+of+earthquake+source+kobe+japan#1222
NCEER Report / Website
State University of New York at Buffalo. (1995) "Report: A Summary of Earthquake Reconaissance Efforts of The National Center for Earthquake Engineering Research Center."
http://mceer.buffalo.edu/research/Reconnaissance/kobe1-17-95/response.pdf
.Composed by the National Center for Earthquake Engineering Reserach Center, this electronic report examines performance of steel, wood, and concrete buildings as well as roads, highways, ports, and harbors. Liquefaction induced lateral displacement is examined as a culpirt for the majority of foundation system failures. A section also discusses the effectiveness of Japan's Building Code based upon the correlation between structural damage and the governing building code for each structure. Finally, social and economic impacts of the Kobe Earthquake are discussed.