Title: NSF/Tokyo Report: Measurement of Lateral Forces on a Sheet-Pile Wall During Seismic Loading Date: 9/19/97 The National Science Foundation's offices in Tokyo and in Paris periodically report on developments abroad that are related to the Foundation's mission. These documents present facts for the use of NSF program managers and policy makers; they are not statements of NSF policy. Special Scientific Report #97-28 (September 5, 1997) MEASUREMENT OF LATERAL FORCES ON A SHEET-PILE WALL DURING SEISMIC LOADING Mr. Eric Liedtke, a graduate student in the Department of Civil Engineering at the University of Texas at Austin, prepared the following report. Mr. Liedtke was a participant in the 1997 Summer Institute sponsored by NSF and the Science and Technology Agency of Japan. Professor Ikuo Tohata of the Geotechnical Engineering Laboratory in the Department of Civil Engineering at the University of Tokyo hosted Mr. Liedtke. Mr. Liedtke can be reached via email at: eliedtke@mail.utexas.edu INTRODUCTION The failure of a man-made river embankment can be a problem for areas lying below the elevation of the river. Seismic activity is one possible source of embankment failure. A method to retrofit embankments to minimize the potential of failure is proposed. The effectiveness of this proposed method is being investigated by experiments performed on soil models. A large shake table is used to simulate seismic events on the models. During an earthquake embankments may fail. In particular if the embankment was constructed on top of a loose deposit of aturated sand, failure may occur due to the decrease in the effective stress of the sand deposit that can occur during an earthquake. The sudden loss in the effective stress of the loose, saturated sand is called liquefaction. The mechanism of this phenomenon is the sudden increase in pore water pressure caused by the rearrangement of the loose sand particles during a seismic event. Thus, in simple terms, during a seismic event the sand and water essentially become a fluid that is incapable of supporting the weight of the embankment. This in turn will cause the embankment to sink into the sand, displacing the sand-water mixture to the sides. A method proposed to reduce the risk of failure of embankments constructed on loose deposits of saturated sand is to construct two sheet-pile walls at the toes of the embankment and parallel to the length of the embankment. The sheet-pile walls will not prevent liquefaction from occurring but might minimize the lateral movement of the soil-water mixture that may keep the bulk of the embankment in place. To study the effectiveness of using sheet-pile walls to minimize the risk of failure of an embankment during a seismic event, several model tests were performed. One such test is described below. MODEL The model consists of a tank constructed of clear plastic 20 mm thick. The dimensions of the tank are: length equal to 100 cm, width equal to 20 cm and a height equal to 40 cm. Toyoura sand was used in the experiment. A dense layer (Dr = 63%) of sand was compacted on the bottom of the tank with a thickness of 12.5 cm. Next, a loose layer of sand (Dr = -7%), 15 cm thick, was placed on top of the dense sand. The sand layers were then flooded and allowed to become saturated. Finally the embankment was placed on top of a thin layer of gravel placed over the loose sand. The gravel was used to minimize capillary pressures in the embankment. The embankment was 20 cm long (the width of the tank), 5 cm high and 5 cm wide at the crest with 2:1 (H:V) slopes. The sheet-pile walls were placed at the toes of the slopes. The dimensions of the sheet-piles were 25 cm tall (placed flush with the surface of the loose sand layer) and 20 cm wide. A different thickness sheet-pile was used for each toe: 1.0 mm and 0.8 mm. INSTRUMENTATION Six pore water pressure transducers were placed in the tank to measure the excess pore water pressures within the loose and dense sand layers during the test. Three accelerometers were placed inside the tank to record accelerations in the loose sand layer. An additional accelerometer was placed outside the tank on the shake table to measure the input accelerations. Ten pairs of strain gauges were placed on each of the sheet-pile walls to measure the strains in the walls caused by the outward force of the loose sand under the embankment during testing. SHAKE TABLE The shake table was 2 m by 3 m and was capable of generation accelerations up to 0.8 g and frequencies ranging from 0.1 Hz to 150 Hz in two dimensions. The accelerations, frequencies and event time were controlled by a personal computer. During testing the table was supported by pressure from oil and two actuators acting in perpendicular directions on the table generated movement. For the test of the model described above the following values were used as input for the shaking table: 10 Hz and 0.34 g for 10 seconds. The pore water pressure transducers, accelerometers and strain gauges were read and recorded digitally at a sampling rate of 200 times per second. ANALYSES The data were collected and are currently being analyzed. Tests on a larger model (2 m x 2 m) will be performed in the near future.