 |
  |
MISSOURI DIVISION |
|
 |
 |
PROCEEDINGS OF THE SEPTEMBER 2000 POST EARTHQUAKE HIGHWAY RESPONSE AND RECOVERY SEMINAR HELD IN ST. LOUIS MISSOURI
EARTHQUAKE VULNERABILITY ALONG EMERGENCY VEHICLE ROUTES BY RONALDO LUNA
MR. NEMMERS: Our second presentation this afternoon is from Ronaldo Luna who is the associate professor at the University of Missouri in Rolla. He's in the civil engineering school and got his Ph.D. from Georgia Tech. He spent four years in the civil engineering department at Tulane before he went to Rolla. Ronaldo is going to talk about is a research project that the University of Missouri, Rolla has with the Missouri Department of Transportation on assessing the earthquake vulnerability along emergency vehicle routes. Ronaldo is part of the team that's doing a vulnerability assessment of the transportation facilities in southeastern Missouri and feeding into the rest of Missouri.
MR. LUNA: Thank you very much, Charlie. I'm very pleased to be here representing the University of Missouri Rolla. This is essentially one of the first times that we have experienced a real cooperation between three state institutions, University of Missouri Rolla, Missouri Department of Natural Resource and Missouri Department of Transportation. Through the Missouri Department of Transportation, we were able to get some funding to get the work done.
I would like to acknowledge some others. Mike Fritz and Tom Fennessey acted as the technical liaisons at MoDOT. I acknowledge all the grad students and undergrad students that were involved. There is a web site that reports earthquakes. Go to http://folkworm.ceri.memphis.edu/recenteqs. The interesting thing to me is they report every earthquake in the past hour, past week and the past six months including the recent event in Batesville, Arkansas. This site is still a constant reminder that the earth is active.
The project or study essentially follows a very classic approach to assessing the hazards of bridge structures. We also looked at flood effects and impact possibilities along some of our highway embankments.
There were two objectives. One was the development of a geotechnical database for future use of vehicle access route evaluations. Subsurface information that often becomes forgotten after the work is performed would be available from our study for future use. So that was one of the first deliverables that was required from the project. The second one was focusing site specific earthquake response of two bridge locations along US 60 route.
Missouri DOT identified two vehicle access roads as being access relief routes to damaged areas in the event of an earthquake. Route 100 is obviously a highly populated St. Louis vehicle route to which we always pay due attention. The other is US 60 that was not supposed to be earthquake prone. The assessment of the two bridges is in the boundaries of one county. A little bit to the east are the St. Francis River and the Wahee ditch. US 60 highway will provide emergency access to the areas from Cape Girardeau down to site of Sikeston in the boot heel. Phase one of this project focuses on the southern portion and US 60 and MoDOT is entertaining a phase two for Route 100.
The borehole sites for the cone penetrometer data were selected by the MoDOT. They deployed their drill rig. There was oversight by the University and Missouri DNR personnel. The borehole data was compiled and reported by MoDOT.
The geophysical field testing allowed them to perform open bore holes and the cone penetrometer work was performed at the University of Missouri Rolla. That information was used to gather seismic data of the soil deposits. The routine laboratory testing that MoDOT does on every single project was performed in Jefferson City and reported in conjunction with the whole. The dynamic soil testing and more specialized engineering tests were performed at the University of Missouri, Rolla. These bridges are lean and unnoticed. There were no large bumps before or ahead of the bridges. However, if they fail, this route would be unusable because there is enough span length in combination with the flooding that would come down from the Wappapello Dam to block the road. We also checked slope stability. Both static and dynamic analysis was performed to assess the likely hood of losing the pavement that gets you to the bridge. Often there is enough height to experience failure in a dynamic mode not in a static mode. Keep in mind, that most of these bridges and roads were designed for static loading alone.
The field and laboratory data was reported by all agencies and the university. We decided to engage in a geotechnical database, making the data a little more usable in the future. We went through a process of design and analysis of the data structure and development of the user interface that was deemed to be extremely necessary. You can always put data in a database but if you can't access it in an efficient way, it will not be used even though it's in electronic form.
The data was calculated and then we analyzed some of the queries. MoDOT is going to do a more in-depth study of the queries of these data. A CD ROM with the data base and the user's manual will be delivered to the MoDOT for future projects.
For a while MoDOT was contemplating many other hazards or sources of information such as geographic sinkholes, ground ruptures, or any particular geologic units or references. At this time MoDOT wanted to focus on information that we are very familiar with. It was essentially the focus of the study to look down through the boreholes for the cone penetrometer data. We are evaluating this data and the data structure to be developed in the phase two. So essentially all the data, both existing bore holes and the new boreholes along each bridge, total more than a thousand. They have been made available through this database. You have core log materials. It follows the same nomenclature, standards, and all the physical properties most DOT personnel are used to. The dynamic soil properties are something new to the Missouri DOT and are handled as a separate and much less calculated data source.
The opening screen of the database is both data entry and a browsing tool. If you want to calculate the data, you start with new structures and then you go down into each of the menus and start entering that information relevant to our highway structure. This can be a box culvert, a bridge or a retaining wall. You continue to find structure information. Very important for the purposes of future use are the location information and any geographic data. This is currently being identified as something of great importance to MoDOT because it does not categorize geographic data with geographic coordinates. They look at data by log mile. When they are trying to relate to other types of data, they have a hard time. They are not used to going out with a GPS and locating themselves within the rest of the world. They just locate themselves within the roadway and if you go down this route, you'll find it. Often this does not lend itself to data fusion with other data products. So phase two will also focus on geographic referencing a lot of work locations. You can access screen two, once you enter some information. The colored buttons will go to calculate core logs, water observations, drain field distributions, or you can add a material or a particular item on a borehole.
Some of the information recently produced at the university is laboratory soil dynamic test results. We have a capability of performing actual tests on soil samples. We're looking to get set up to have essential elements to gather information that is more low strength nature. We don´t want to rely only on field data to get those elastic parameters that are received from low strength measurements. This is the low displacement strip, essentially three different samples for different strokes. You can see the degradation of the modulus as you decrease the stroke. It's showing that soil groups are strength dependent. If we plot the strength versus the shear modulus, we can see the large degradation to over half a module magnitude.
Defining the soil properties is the precursor to move on to some of the most prevalent ground ruptures that are expected in the boot heel area. The boot heel area of Missouri has one of the largest deposits of liquefaction susceptible areas in the United States and close to the world. However, we haven't had a recent earthquake to experience that. In a recent travel day, I went out to southeastern Missouri to meet with folks in Portageville, a very small, populated area in the boot heel. They have generations in their families that can identify where the liquefaction sand boils are. These are prehistoric events or they were not recorded in history. There were probably 18 consequences. At the locations where the sand boils or the sand liquefied came up in a dike to the surface, their crops had a very low yield at a particular time of the year. If you take aerial photography, you will be able to identify liquefied, soils. At that time there were not many bridges or structures around to be impacted. When we were driving and visiting some of the sites, we could see the sand boils from the "Hill". The "Hill," what hill? The highway embankment approaching the bridge is called the hill. It's pretty flat there.
We have a classic test to assess the liquefaction effect. First you propagate the earthquake and the motion from what we have identified as something close to being bedrock. You propagate that up to the surface or to the particular layer that you're trying to evaluate and we typically use a shake-type analysis. We can then calculate the liquefaction potential. You can either use cone penetrometer data or soil penetration test data or go to a more rigorous liquefaction modeling technique. You can do a deformation analysis of the earth structure in combination with the abutment. Finally you can anticipate some of the foundation failures for the bridge you are studying.
This is a combination of the shake and liquefaction analysis using the simplified procedure. You can see at about 50 feet is what was identified as being a reasonable hard-firm layer that we could label rock. I know that some geologists would question that but we reached refusal a number of times and called that bedrock or very hard soil. We applied ground motion. There are challenges in evaluating ground motion and doing earthquake hazard assessment in the Midwest or in the boot heel area. We do not have strong ground motion recorded time histories to perform analysis. They were not available and in 18 events, there were no accelerators. We have a lot of very small earthquakes and what we would like to look at is the big one. There is very little history and almost absolutely no data. So we have to go through an exercise called developing a synthetic ground motion. The Mid-America Earthquake Center in Illinois has a task of developing some of these ground motions with a number of hypothetical events. The locations they have chosen are St. Louis, Carbondale, and Memphis. All of these are pretty far from US 60, so we are faced with the challenge of developing yet another synthetic ground motion for both of those bridge locations.
There is currently a lot of discussion as to what ground motion we're going to use and how we're going to scale it. Because if you place in the map these three ground motion locations, you cannot really do a linear interpolation and predict the ground motion because all of these sites are away from the epicenter. We expect there will be stronger ground motion near Sikeston and the boot heel area. This looks at low liquefaction, issues of quality, the focus is not around the epicenter, and a perfect reflection of what happened in Watsonville as pointed out before.
We are not developing any new ground motion but we looked at the three available ones. In phase two we are planning to look at that in more detail and capabilities in house to develop this synthetic ground motion. In California they have the luxury of developing ground motions that are based on data. But we have to come up with assumptions to generate a synthetic time history.
A this location we input acceleration which is .15 at the base and as it moves up the soil column it degrades and amplifies a little bit more as it reaches the clay. It degrades through this sand and at the surface it reaches .15 again. So it degrades and then it amplifies again when it reaches the topmost deposit. You can see the flow graphs reflecting some of these sound waves and some data that was lost here. We have penetrometer data to verify this and to fill in the data gap. We then make a comparison of the seismic stress ratio and the seismic resistance. The ratio of these two is going to be the factor of safety that will define when liquefaction occurs at that particular elevation. In a couple of locations, we are getting very close to a factor of safety of one. In some more intense accelerations, the ground does liquefy at different locations in profiles. So the ground does move if you make it move. It's very prone to liquefaction throughout the area. So the next step is seeing how the bridge is going to respond and that's what the seismic condition and assessment of the bridges is.
Global performing criteria and the engineering performance criteria was defined in the analysis and it was done via a computer model using a site 2000. This was a one-year project. We were out in the field about six to seven months collecting data. By the time the computer model was ready to look at the ground motion, there was not enough data to do the soil structure interaction and look at the combined effects of the bridge and the soil that will add damping to the response of the earthquake. The examples that we were able to run and deliver some preliminary results to Missouri DOT were without including soil structure interaction.
There are two bridges at the St. Francis site. This bridge is the old bridge. You can see that the abutments or the supports are on a pile foundation and pile caps. There is some fill here, the abutment, and the superstructure. These are not integral design. The abutment is separate from the bridge. However, parallel to this bridge, when US 60 was enlarged, a new bridge was built that has totally different design criteria. This is an integral type bridge where there are essentially no supports. There is a strong connection between the abutment and the superstructure. However, we looked at the old structure and we're moving to look at the new structure. Our approach was let's look at the one that will most likely fail. This is an example of the modeling exercise. This finite elements mesh has over 600 elements and about 70 shell elements. This example is showing the outlook of the fractures along the longitudinal plane. There were some spring constants applied at these supports. They do not yet include the soil structure interaction at the supports of the foundation.
This bridge is built on a skew. When you apply the ground motion in the longitudinal direction that skew actually induces, loads of about 36 percent in the transverse direction because of the difference in the length of the structure elements and also along the girders. This was one of the preliminary findings and we have yet to look at three more bridges for this site.
This study is essentially just an exercise of data development. The future is to go to GIS and try to gain from the site-specific data and do a more comprehensive analysis. We should not look at just one bridge. We want the entire system to work. We want to extrapolate from this intense data development efforts of the state agencies and to share them and be able to have them in more large-scale and more small-scale GIS map.
These site soils do not liquefy for an earthquake probability increase of 10 percent in 50 years. However, when you increase the earthquake to a probability of 2 percent in 50 years, soils liquefy with the ground motions used. This ground motion is going to be your driver. I want to be careful to mention that these ground motions are all hypothetical. We need to have something to go by and we think they are possible ground motions. Seismologists will debate this with us all day. They have no data. It's just an analytical model. For bridge abutments, large displacements, and pile rupture may be experienced during a large event. Once liquefaction occurs, the pile supports are not designed for any lateral loading. We do have battered piles. Those battered piles were designed for lateral loads but not for those of an earthquake. Slope stability analysis shows the same incidence as liquefaction analysis was at probably 10 percent. There were no slope failures however for a probability increase of 2 percent in 50 years, the analysis shows a strong likelihood of failure. Without soil structure interaction, fixed pile caps, the bridge were not extreme and skew of the bridge induces transverse moment of about 36 percent. The foundation is not being combined with the bridge and there is no soil structure interaction effect. This is essentially to define our procedure to proceed with future sites and establishing good working relationships between all the agencies. This type of study cannot be done by one, two, or three engineers. It really requires a number of people working together and an extremely good manager like Neil Anderson to get results. In the future work we will be working with Bob Berman to get some site-specific strong ground motion at St. Louis University. He's a very well known seismologist in this area. For determining the liquefaction potential in response to ground motion, we're looking at a number of things. We want to model soil columns in the boot heel area to go beyond the simplified method using some effective stress and finite element analysis to find the difference. We will look at the behavior of the soil column as opposed to just applying factors of safety, will it fail or not. Soil stability is really dependent on the geometry of the site. Most of the sites are going to have similar soil properties, depending on when the cuts were made or when the hills were made.
The abutment and the structure stability are the two items that really raised some flack in the Missouri DOT. What will happen to my bridges. The Missouri DOT is aware that the bridges are very vulnerable. They were not designed for dynamic loads. A lot of them are very old and even though this project started as a geotechnical type problem, MoDOT wanted to include structural components. As the project has matured, a lot of attention is being given to ground motion and deformations of soil. If the Wappapello dam fails, flooding would cover the entire area and those tributaries crossing US 60. This is based on a Corps of Engineers study of 1985. There were field visits to the dam and the different man-made ditches and natural drainage patterns to confirm some elevations and that vulnerability is still there. This is phase one. We intend to go on to phase two shortly with some additional initiatives that MoDOT is interested in looking at such as liquefaction studies and larger studies of the Mississippi. That's about all I have.
MR. NEMMERS: Thank you very much. You indicated that you're doing some global monitoring of the St. Francis Bridge. Are you instrumenting the bridge at all?
MR. LUNA: No, we're going to use global positioning systems to geo reference the boreholes.
FROM THE AUDIENCE: MoDOT is planning on doing some instrumentation for the new Emerson Bridge there at Cape Girardeau.
MR. LUNA: That is a more worthy bridge.
 |
Send mail to scott.holder@fhwa.dot.gov with questions or comments about this web site. |
