Earth retaining structures constitute a vital component of the civil infrastructure across the U.S. In high seismic risk zones, these structures are occasionally subjected to strong earthquakes that can threaten their integrity. Research is needed regarding seismic demand and performance of these structures under strong ground motion to improve existing design procedures. Experimental research of this type has traditionally been conducted by testing small physical models on actuator-driven shake tables or in geotechnical centrifuges. Small model tests give valuable information on general behavior but represent a compromise with respect to field structures because of low stress conditions (if 1-g), scaling effects, idealized backfill soil, idealized compaction conditions, and the need for non-standard structural elements, such as reinforcement and facing elements. What is needed for better understanding of the performance of reinforced soil walls during earthquakes is full-scale testing. For the first time, such testing is possible using new shake table equipment that be available at the University of California-San Diego (UCSD) starting in late 2012.
Current seismic design procedures available to state DOT’s (e.g., AASHTO) may be excessively conservative for many wall types, but also may be missing important design considerations important for good seismic wall performance. If current design models can be better understood or better design models can be developed that efficiently reduce conservatism while accurately capturing the seismic design issues that warrant greater attention, there is significant potential to reduce overall wall costs. This is important for existing use of MSE walls and future use of MSE abutments, which is a relatively new and cost-effective application for this technology.
The new AASHTO seismic design requirements (moving from a 500 year to a 1000 year design earthquake) have increased the seismic demand on walls – the biggest impact is on the width of the wall design section now required to resist the seismic loading. Significant savings in wall costs could be achieved if the proposed research would allow for reduction of this width. Reducing the width of the wall section also directly reduces the amount of shoring and excavation needed to build a wall, resulting in significant additional savings. Increased demand on the internal components of these wall systems will also occur, and developing a more accurate methodology to estimate the internal stresses in walls is also needed.
A project, funded by the National Science Foundation (PI: Fox; total funding: $550,000), is currently underway at UCSD to build and test prototype full scale MSE walls (7 m, or 23 ft., in height) using UCSD’s Large High Performance Outdoor Shake Table (LHPOST). In addition, another project for Caltrans is in the final approval stage to use the same facility to conduct full-scale tests of “true” MSE bridge abutments (i.e., shallow foundation directly sitting on MSE wall). The purpose of this current pooled fund solicitation is to extend these projects by performing numerical studies and one or two additional full-scale MSE wall tests on the UCSD shake table. This work will complement the existing projects and allow us to develop a more complete understanding of the seismic behavior of reinforced soil walls without bridge abutment loads. Opportunities to use large testing facilities such as the LHPOST are rare, and the ability to take advantage of such facilities with the majority of funding covered by other agencies is even rarer. Thus, the cost of the proposed project is comparatively low because we can “piggyback” on other existing projects.
The objective of this project is to perform numerical studies and use the LHPOST to investigate the dynamic performance of one or two full-scale (7 m) reinforced soil retaining walls constructed using realistic materials and methods. Considering that these walls will be substantially taller than for any similar previous research (by a factor of 2), a key focus of the proposed research will be on the influence of wall height on overall system response (i.e., stability/deformation) and the distribution of dynamic tensile forces (i.e., seismic demand) in the soil reinforcement. Other focus areas will include dynamic earth pressure on facing elements, effects of dynamic loading on soil-reinforcement stress transfer mechanisms, and permanent deformations after dynamic loading.
The tests will be conducted using a unique large soil confinement box (LSCB) that is currently under construction as part of a recently funded NSF grant. The scale of these tests will permit wall construction using realistic soil types, compaction methods, and structural elements. The box will also have a unique design that permits different boundary conditions at the rear of the soil mass, including a water-filled bladder or geofoam layer.
Scope of Work
The project will consist of dynamic testing of one or two MSE walls, each approximately 7 m in height. The specific types of walls to be tested will be established in discussion with the TAC (Technical Advisory Committee) for the study and representatives from each of the funding organizations (e.g., state DOTs). The wall specimens will be constructed using realistic soil types (e.g., well graded granular material with some fines), compaction methods (e.g., rolled soil lifts), reinforcement (e.g., geogrid or steel strips) and facing elements (e.g., modular blocks or precast panels). Each wall will be shaken at several heights during construction (e.g., 2 m, 6 m, 10 m) to assess the effect of wall height on dynamic response. Tests on a given wall at intermediate stages will be conducted using only moderate excitation. Once construction is completed for each wall (i.e., full height), specifications for the LHPOST indicate that peak accelerations of approximately 0.8g can be achieved.
The LHPOST is located at the Englekirk Structural Engineering Center of the Powell Structural Research Laboratories at UCSD. The LHPOST is the second largest shake table and the only outdoor shake table worldwide. The table measures 7.6 m × 12.2 m and permits simulation of large earthquake ground motions. In its current configuration, the LHPOST has a stroke of ±0.75 m, a peak horizontal velocity of 1.8 m/s, a horizontal force capacity of 6.8 MN, and a vertical payload capacity of 20 MN. The testing frequency range is 0-33 Hz. Although designed for full 6 degree-of-freedom (DOF) capability, the LHPOST currently has uniaxial (horizontal) motion capability. Work is currently underway to upgrade the table with vertical motion capability.
The MSE wall specimens will be instrumented with sensors, including accelerometers, strain gages on reinforcement, earth pressure cells, displacement transducers and possibly shape acceleration arrays (SAAs). SAAs (Abdoun et al. 2005) are a rope-like array of sensors and microprocessors that fits into a small (27 mm ID) casing. If the budget permits, SAAs will be installed and grouted in vertical drilled holes after backfill construction (e.g., near facing, at middle of reinforced zone, at back of reinforced zone) and will measure the change of casing shape similar to an inclinometer. The advantage of the SAAs is that they can measure lateral displacements in real time during dynamic loading. All other instrumentation is currently available through NEES@UCSD at no cost to the project. Sensors will be placed within wall specimens as needed during construction. Strain gages on reinforcement (geogrid, soil nails) will be load-calibrated to the fullest possible extent to reduce measurement variability. Total pressure cells suitable for dynamic pressure measurement will be placed behind the facing, similar to methods previously used by Dr. Elgamal (Wilson and Elgamal 2008, 2009a,b,c, 2010). Video footage of the top surface and front face will also be collected during testing. Similar instrumentation methods were used in a full-scale study of the static stability of a 55 ft. (17 m) high retaining wall conducted by Dr. Fox (Runser et al. 2001). Excitation for the full-scale walls will consist of a range of input motions including low amplitude frequency sweeps and records from one or more large earthquakes.
In addition to the experimental work, numerical studies will be conducted on dynamic performance of MSE walls using the geotechnical analysis software FLAC, which has a long track record of success for soil-structure interaction problems and is available to the project at no cost. Investigated variables are expected to include wall height, facing element types, reinforcement types and layout, soil backfill properties, and shaking time-histories. FLACTM (Fast Lagrangian Analysis of Continua) is a leading computer analysis package for deformation and stability of geomaterials developed and supported b
THE SCOPE OF WORK AND BUDGET HAVE BEEN MODIFIED BASED ON OTHER ASSOCIATED WORK THAT WE WILL BE PIGGYBACKING ON TO ACCOMPLISH THESE TASKS.
Minimum state commitment is $10,000 per year for two years (total of $20,000 per state).
Since this is intended to be a continuation of the ongoing NSF grant, the principal investigators would be the same as for the currently funded project:
Patrick J. Fox, Professor (PI)
Ahmed Elgamal, Professor (co-PI)
Department of Structural Engineering
University of California-San Diego