 |
  |
MISSOURI DIVISION |
|
 |
 |
PROCEEDINGS OF THE SEPTEMBER 2000 POST EARTHQUAKE HIGHWAY RESPONSE AND RECOVERY SEMINAR HELD IN ST. LOUIS MISSOURI
TURKEY EARTHQUAKE REPORT BY JIM COOPER
Thank you very much, Ed. I want to take this opportunity to go over a couple of earthquakes that occurred last year. First of all, I want to thank the co-author, Hamid Ghasemi, who had to go to Japan and Taiwan to talk about the Turkish and Chichi earthquakes.
We're looking at Istanbul at the Straits of Bosporus, moving over into Europe, Bulgaria, Hungary, Poland, and then moving to the east; Syria, Iraq, Iran; the capital of Turkey, Ankara. The mission was to look at the damaged structures on the North Anatolian Fault Zone.
The Trans-European Motorway is a very crippled transportation route to the economy of Turkey. I discuss some of our key observations, sample of damaged bridges, recommendations, and conclusions.
We were at the locations of the two epicenters occurred. August 17, 1999, the big earthquake occurred and resulted in significant building damage, very poor quality construction. The construction associated with the Trans-European Motorway and other highway structures and bridges in Turkey that we did not investigate were actually constructed to significantly higher tolerances.
There was between 30 to 45,000 casualties. Many unreported are unknown because of immediate burials by family searching through the rubble, a real tragedy.
They had 45 seconds of strong ground shaking in the first earthquake. It was followed a couple months later on November 12 by a 7.2 magnitude earthquake located to the east of Iznik and Kocaeli, in the Bolu/Duzce Region. A 7.2 magnitude resulted in more than a thousand casualties. There were some very interesting and significant damage to bridges under construction on the Trans-European Motorway.
Hamid Ghasemi and Phil Yen are two young engineers in my office who went on this trip with me. It's very important to have succession planned. So this was an outstanding opportunity for some young individuals to go along, learn the business by getting firsthand exposure and experience. Roy Imbsen is a formal CALTRANS employee, president of his own company, and part of the team. We also have a member from CALTRANS, senior structural engineer, Saad El-Azazy.
On the first visit, we spent about 30 days in Turkey; the second visit about ten days. We had quite a bit of exposure to what is there. The mission on the first visit, after the August 17 earthquake, was not so much reconnaissance of looking at what happened. It was to help the KGM that is the Turkey equivalent to Federal Highway Administration Engineers and their consultants in helping to plan for restoring the damaged system from the first earthquake.
Here is a map superimposed by the Trans-European Motorway, the second earthquake, the Ducze earthquake, then the first earthquake and Istanbul. There were three major segments of construction along the Trans-European Motorway going from Istanbul to Ankara.
The first segment in this area, roughly the epicenter of the first earthquake, was basically under the design and construction supervision of the French and Germans. These structures were designed in the 1960s and constructed in the 1970s. This is kind of important because it gives the status of seismic design for bridges was in those eras.
From basically the Kocaeli/Iznik area over to Ducze was the second stage of construction. This was designed in the 1970s and constructed in the 1980s. This was under the supervision and direction of Italian designers and contractors.
The third section and the most recently completed section is from the Bolu area down to Ankara. It was designed in the 1980s and constructed in the 1990s. It was under design and construction of U.S. contractors in support of the Turkish government.
As bridge construction started in the 1970s and 1980s, seismic design was not quite as well understood for highway bridge construction. The Trans-European Motorway is completed for approximately 1500 kilometers. The Turkish government plans to complete an additional 800 kilometers.
The Motorway is very, very important to the Turkey economy. Much like the transportation network in our county is very important to local economies as well as national commerce.
Superimposed over the system is the North Anatolian Fault Zone. The movement of commerce and goods is critical. Out of a hundred percent of movement of goods and services, three-quarters of it is related to international commerce that is absolutely essential to the economy of Turkey. About one-fifth is related to domestic commerce. This is moving along the Trans-European Motorway. Only about two and a half percent is moved by rail and sea. So the highway system is the primary movement of goods.
Let's take a look at the tectonic setting. Here you have the Asian plate up, Istanbul, the Straits of Bosporus, and the North Anatolian Fault Zone. The Asian plate is moving in a southeasterly direction squeezing up the Anatolian plate that is moving at a northwesterly direction. You tend to get a right lateral displacement and movement on a fault zone. That happened in both the August and Noember earthquakes.
Professor Barka and several others in the U.S. and Turkey have been looking at the seismicity of Turkey. The North Anatolian Fault Zone and the capital of Ankara is where the action was. But in recent history, if you look at 1939, '42, '43,'51 '44, '67, you can see a progression of seismic activity moving towards the west to Istanbul. This is of significant concern.
Several Turkish professors, as well as some folks with the U.S. Geological Survey, have predicted and still predict the movement and migration of the plate boundary action to the west towards Istanbul. The first section of Trans-European Motorway is a concern because there were no seismic design requirements.
Looking at the earthquakes, the numbers in the white boxes represent recorded acceleration levels from the August 17 earthquake. The numbers in the yellow boxes represent acceleration in percent of gravity from the Ducze earthquake, November the 17. You'll see that one record here was clipped at 1-G, 100 percent of gravity and one off scale. For the August 17 earthquake, we talked about several. It was discussed yesterday about the effects of distant areas amplification. Here are some effects of some amplification relative to the epicenter. You get up to 37 percent gravity and 41 percent gravity at pretty significant distances and somewhat attenuated moving to the west.
The acceleration record from the first earthquake tells a big story. This is roughly the 45 seconds of strong ground shaking from the earthquake. Here you have ten seconds of pretty strong ground shaking quieting down the seismic activities. The structures, buildings in particular and bridges to some extent were softened from the first jolt. Then 20 seconds later, another good five to ten second jolt and this is what did a lot of damage from the first earthquake. The maximum acceleration levels were about three-tenths of a G. It was a little higher than three-tenths of a G recorded at so-called Lighten T Station in the epicenter region.
We looked at the second earthquake. It produced an entirely different energy pattern. You can see quiet activity. You can see a real big jolt then quieting down. It was about eight-tenths of a G horizontal acceleration maximum.
On our first visit for the first earthquake, we were charged with looking at structures in this basic area of Turkey and only on the Trans-European Motorway. There were others that looked at structures on their national or state highway system. There was a total of about three or four bridge collapses total. We only saw one.
Turkey uses a combination of Euro codes and AASHTO Standard specifications. Today, they're using the more current AASHTO design codes. I have a comment on the application of the AASHTO design codes for bridge designs that are somewhat complex. Misinterpretation of these codes occurred and some difficulties arose in the field as a result of that.
There were a lot of surface fault rupturing or surface traces of faulting activity. This crossed nearby structures in the earthquake adjacent to and underneath. The second earthquake, further to the east, crossed underneath one of the new major structures that was 95 percent complete.
A large part of the damage from the first earthquake was the failure of restraining mechanisms-shear keys. There are one or two more restrainers, bars, and cables that provide longitudinal continuity. There are also transverse shear keys that are typically concrete blocks that lock the girders in from having transversal movement. The shear keys failed.
As we went east, we moved toward the newer design philosophy where the more recent work was done and the more seismic resistance was designed and built in. As a generality, in spite of the tragedy and severe damage, bridges performed relatively well. There was relatively good construction and relatively good inspection. That's the difference between the performance of the transportation network in Turkey and the performance of buildings for which is another very interesting study unto itself.
One totally collapsed bridge we saw was at the Arifiye overpass, less than 50 kilometers from the epicenter. A span dropped off the abutment. This abutment is very interesting. Mechanically stabilized earth wall, the first real good test of this type of design technology and construction. Then with the drop of the span, the fault zone actually went underneath the drop span, resulting in a domino type of collapse. We actually did not see this. This was all cleaned up within about three to five days because of the importance of the Trans-European Motorway. It was imperative to keep commerce moving.
Mechanically stabilized earth wall performed very well. There was a drainage system that went underneath this wall and you can see some sloughing and settlement of the wall that is very important. A lot of folks around the country are using this type of construction including the seismically active areas in California. This presented a great opportunity to study the performance of this type of construction that did relatively well.
I want to show you what surface features can look like. This was a poplar grove, which was probably about 400 meters from the site of the Arifiye bridge collapse. Here you have rows of poplar trees planted ten meters apart and the spacing between trees is two meters. What you see here is about 25 meters of actual fault displacement that´s 400 meters or so from the bridge site, very visible. You can see this offset and a whole row of poplar trees displaced some three meters to the right looking across the fault trace.
When people talk about earthquake proof, it drives me crazy because nothing is earthquake proof. We can make things resistant but there are no guarantees.
Here´s another example of fault movement at the Golcuk Naval Base. This is where the Turkish Navy lost a significant number of their officers. It´s about a four-meter offset.
The Sakarya viaduct is one of the structures on the Trans-European Motorway, ten-span structure, two spans continuous for live load in the deck, three spans, then repeating for the rest of the ten spans. Significant lateral offset. If you look down the fascia girder line here, you'll see an offset of about six-tenths of a meter.
We worked with the Turkish engineers on how to get this back into service as quickly as possible.
Why do we have two parallel structures with such significantly different responses? First, these structures were built at different times; and secondly, the shear keys are what failed in this structure. There was better quality control with the shear keys of the eastbound span and consequently the difference.
The structure itself was designed to two-tenths G static lateral coefficient in 1975. AASHTO shear keys proved to be ineffective and sheared off. We would call this minor to moderate damage but it was a major catastrophe to the Turkish government because of the inability to respond quickly.
They got the debris cleaned up quickly. But to move in and do the engineering work, the repairs, the rehabilitation, and get their contractors in place was another story. I think that's what this meeting is all about: talking together because there's a lot that needs to be done before the event as well as after the event.
I want to focus on is the shear key. This is the actual end of one of the box girders on its end. There are cutouts. These are the key ways and this is how they chose to provide the lateral restraint with just concrete blocks. They just crumpled and they actually fit into the shear key. So when you get the shaking of the ground, you get the shearing off of the top of the concrete shear block. Your whole structure displaces and it's free to flop around in the seismic activity. The first earthquake had the first jolt, which softened the structure, and the second jolt is what did the significant damage. You can see the offset from the elastomeric bearings.
It's always nice to use white lines on bridges to measure offsets. This helps the seismologists and geologists to see the offset at the end from the approach slab, which was damaged typically.
If you looked at overpasses on the Trans-European Motorway, you'd say, "Hey, these look pretty much like ours," two-span over-crossings, three-span over-crossings, four-span over-crossings, pre-cast, pre-stress concrete construction, very typical and prevalent in Turkey along the Trans-European Motorway.
The type of damage very is similar to what we've seen in California earthquakes where you have the mass of soil behind the abutments. You see the cracking of the wing walls, much too similar to what we saw on Northridge and San Fernando and other earthquakes. Its damage that is difficult to control, it will happen. It is not considered major damage. It´s a headache and a pain in the neck for the maintenance forces.
Some typical underpasses are basically concrete box structures with wing walls to restrain the sloping surface of the built-up roadway on the Trans-European Motorway. You get a lot of separation, discontinuities, cold joints, the wing wall, and the roof of the concrete box. Literally, we had a culvert structure. There was also a settlement halfway through the underpass. This created a natural expansion joint about halfway through.
The second earthquake hit the newer section, the 24 kilometers of the Trans-European Motorway still under construction that will connect Ankara to Istanbul. There were two major viaducts. The first viaduct suffered no damage in the first earthquake. It was 95 percent complete with some 59 or 60 spans, 2.3 kilometers in length. It was a major structure.
The second viaduct was only 5 percent complete. They were in the process of constructing the foundations and that led into a major tunnel 3.3 kilometers in length. It was about two-thirds of the way through excavation.
With the first earthquake, which was located over in the epicenter, there was no damage to any of these structures. The second earthquake caused significant damage to the viaduct that was 95 percent complete. They also had to close both tubes in the tunnel.
On the first viaduct that was damaged by the second earthquake, we happened to inspect it after the first earthquake so we had a real good baseline on which to measure the results.
This was completed portion of the Trans-European Motorway but it was not opened to travel. Landslides took out traffic. This was actually a cut that was made for the new construction. So land sliding was prevalent and it was something that needed to be watched.
Here we´re looking at what we call Viaduct No. 1, the main structure. It was designed to "modern AASHTO codes," and it incorporated energy dissipation types of systems, something we're encouraging in this country. Those systems really got beat up.
I want to show you what a concrete box girder looks like. They're 40-meter span lengths and pretty significant concrete box girders. These are the box girders that were lined up waiting construction of Viaduct No. 2. Typically, you don't store box girders like this.
This is a result of the surface waves and you get that good shock and its pretty significant. I don't remember the weight of these girders. But tip them all over on their side and you can look at kilometers of lines of girders laying on their side toppled just like this.
Here's the Viaduct Structure No. 1. This ten-span unit is typical. These ten-span units repeat themselves. These E's don't represent expansion joints; they represent energy dissipation units at the top of the piers. At the top of each and every pier, there's an energy dissipation unit where the expansion joint exists at the center of each ten-span unit. Along this 2.3-kilometer viaduct, there are pistons that allow and accommodate temperature movements and displacements caused by temperature. At each end of the ten-unit spans, there are restrainers or cables. It's kind of a three-prong system. When we were at the site the first visit for the first earthquake, the designers assured us that there was absolutely no need to provide restrainers to the energy dissipation units as well as the pistons that controlled total deformation. As it happens, this was the savior of the viaduct, these little simple cable restrainers.
If we look at the energy dissipation units that were on the pier caps, there was no damage. Here's the piston, which controlled temperature displacement. There's a series of ring devices. They can be very simple or very complex depending on where they're located in the structure. The actual elements, the crescent moons, or the C elements will bend and absorb energy. This is just a template or holding plate to hold the energy dissipation unit together.
From the first visit, there was no damage. This is the typical type of bearing that was in existence on the structures. If you look up on the pier cap at one of the expansion joints, you see the complete system with the hinge restrainers.
The Ducze earthquake came along November 17. At this bridge, the columns were actually rotated about 12 degrees as a result of the seismic action. This was the only rotation of columns that we noted. Unfortunately as of Friday last week, there has not been an excavation to determine the condition of the pile foundations and whether there is a damaged area along the structures.
If you look underneath the column looking up and carefully follow the line of girders, you'll notice the pier table. The pier cap is not exactly 90 degrees. It actually rotated some 12 degrees.
Bearings were pretty well beat up. The pot bearings kicked up and got displaced. There´s pretty much a mess down underneath the Teflon stainless steel sliding plates. To my knowledge, our team has been the only folks allowed to go up on top of the pier caps. Others were told it was too dangerous.
For some of you older folks, you may remember seismoscopes. They are little smoke glass plates that in an earthquake provide a little tracing of seismic motion. Low and behold from displaced bearings and Teflon plates, you can see scratches or the trace of the displaced, dislodged plates.
This was very typical on many of the displaced bearing seats we saw. From these we deduced what happened. A little rumbling to begin with, then precursor waves, then those big shakes, which twisted it. You've got the mass of your girders hanging up and you've got your substructure that's moving significantly and kicking out and consequently, producing this type of telltale evidence.
We have an important lesson here. The bearings were designed by one person and the energy-dissipating unit was designed by another. The capacity of the EDUs was 480 millimeters of displacement. The design displacement for the bearings and the bearing seats, actually, was 375 millimeters. So you have a mismatch of designs. The lesson being the structure, the bearings, the restrainers have to be designed as a system.
Energy dissipating units are interesting. The energy dissipating units never had a chance to work. You'll see a bunch of destruction. This is a piston on the ground, a piston that was sheared off from its table on top of the pier cap.
The connections are also important. These connections were not designed. They were hit with this one big shock, the second earthquake, and they failed and rendered the whole system inoperable.
Had it not been for restrainer cables, this structure would have been flat on the ground. It's bad enough as it is and very difficult to take care of and nothing has been done to this date to repair the structure.
This is the hinged joint. It's not complete. The expansion joints were not in place. The structure was 95 percent complete at the time and pretty significant damage.
There´s a different kind of shear key here. Shear keys were actually placed between the box girders. They did their job in terms of horizontal restraint and did fairly well. But they were pretty well beat up from the impact of the whole pier moving back and forth.
General conclusions. Important bridges. From the first earthquake, the system really did perform well. We made some recommendations. The Turkish Federal Highway Administration needs to be concerned about larger earthquakes in the Istanbul area according to work done by others. It's interesting to note the adoption of AASHTO criteria for design.
The AASHTO specs are difficult to apply and there can be different interpretations. When different interpretations are made, design is done incorrectly and you can run into problems.
Obviously, there will be future events and varying performance. Most of the damage that we saw was a direct result from surface fault rupturing and shaking. Liquefaction was evident. I want to go very briefly to the Bolu Tunnel. We had the opportunity to visit the Bolu Tunnel on our first visit while it was under construction. They had geologic problems because the geology varied throughout the 3.5-kilometer length. They started out with three tunnel lining sections. They had to modify the design to accommodate some very soft squeezing clay. That's the area we really ran into problems from the second earthquake.
The twin bores, 17 meters wide, accommodated two lanes of traffic in each direction. The New Austrian tunnel method was used whereby pilot tunnels were actually bored prior to the main excavation of the tunnel lining system. So the boring took place and excavation was completed in segments. The face of the tunnel provided sets of reinforcement and stabilizing types of construction techniques were used to construct the inner lining. The excavation was done then everything was moved ahead into the next segment.
There wasn't any seismic design associated with the Bolu Tunnel until 1998. It certainly performed well in the first earthquake. We made some recommendations that were actually being considered at the time of the second earthquake and being implemented.
We suggested additional instrumentation to be added but they didn't have additional instrumentation at the time of the second earthquake. But at the time of the second earthquake, we had closure of both tubes from the east port of the Elamalik Portal about 300 meters inside.
Interestingly enough, it petrified the workforce. The workforce refused to go back to work whether it was Viaduct 1 or within the tunnel. So they had some psychological risks. Actually, those people are probably smart for staying out. Some of us are not so smart. On December 3rd, I was there and the last person I know foolish enough to go in.
Here you have the sloughing. This is the ceiling, the roof area. You can see the sloughing that came in but you just have to say you saw it. We did take some measurements. There was spalling between sets of the shotcrete and the hinge areas. Here´s a partially completed tunnel liner segment with the waterproofing, the interior liner, the intermediate liner, and the interior reinforced concrete liner. There's drainage underneath and there are utilities.
As a result of this particular earthquake, and the second earthquake, they are actually going to relocate a portion of the 3.5 kilometers.
The one lesson that we really learned is where you have a release of energy, you get this big push, then the near-source zone, it doesn't attenuate. Your structures really have to absorb that energy right near the fault zone. Of course, it's difficult to predict where that fault zone is. That's the problem for the seismologists, geologists, and for the design engineer in trying to accommodate it.
Multidisciplinary approach, that brings us to conclusion of what this meeting is all about. Over the past 40 years, we've come a long way from a scientific and engineering point of view in understanding how structures behave.
Every bit as important, if not more important, is the effects of public policy and decision makers and how much can be provided to protect the investment. Legal and financial strategies need to be looked at. And I think in looking at the disciplines of all people involved in this type of meeting.
Awareness of the earthquake problem is important if you want to get anything done. I can remember coming to St. Louis in the late 1970s and putting on a training course for the Missouri Department of Transportation. We got laughed out of Missouri. It was a great sleeping effort for those folks at that time because they just weren't aware of the earthquake potential. They didn't believe there was a threat.
Over the last 30 years there's been quite a bit of awareness, information dissemination, and meetings like this. Missouri DOT has taken the problem seriously. Illinois DOT has taken the problem seriously. My friend Ed Wasserman in Tennessee is very serious about it in the Memphis area.
Finally, when we bring it all together with the discipline of many of you folks, plan, plan, plan, and plan to respond.
So with that, I thank you very much.
ED GRAY: Jim talked about when is the next earthquake. We're always asking that. I'm a historian by trade. The historical repeat interval gives you a much better way of looking at when we will have the next major earthquake. A 6.3 earthquake is about equivalent to what happened in Missouri on Halloween 1895 or 105 years ago. The mean repeat interval is about 55 to 85 years. Based on this, we're overdue on that size of earthquake. 7.1, last one was probably one of the 1,874 earthquakes that was the part of the New Madrid series. That was 189 years ago.
The 7.6-earthquake is what Dr. Otto Nutley said is about the estimated strain that's built up along the New Madrid fault. It is our planning basis at Emergency Management here in Missouri. So we're right close to entering that period.
For greater than an 8.0 earthquake the average interval between earthquakes ranges from 450 to 650 years. We don't know which end of that interval we're at. We could be at the end of one or we could be at the start of another.
Our next speaker is Steve French. He's a professor of city planning, director of the Center for Geographic Information Systems at Georgia Tech. He works with the Mid-America Earthquake Center at the University of Illinois at Urbana-Champaign. He's done research on the most recent earthquakes, including Northridge, Loma Prieta, and the Whittier Narrows. He's a member of the Earthquake Engineer and Research Institute.
 |
Send mail to scott.holder@fhwa.dot.gov with questions or comments about this web site. |
