Justin Kindt and Casey Jones present the results of a cross-hole seismic tomography survey in the downstream test section at the Tuttle Creek dam modification project
The Tuttle Creek dam was constructed in the 1950’s to create a reservoir for flood control on the Big Blue river. The dam is located approximately 4.8km north of the city of Manhattan, Kansas, US. In 2005, a contract for seismic remediation was awarded by the Kansas City District Corps of Engineers to Treviicos South (the general contractor). The planned project was to include the construction of a 1920m long jet-grout cut-off wall below the upstream shell of the dam embankment. A test section was installed downstream in the campground area to the southeast of the existing dam prior to the construction of the main cut-off wall. The test section project involved the installation of nine groups of three jet-grouted columns and nine groups of three deep soil mix columns. Each group of columns was constructed using different mix proportions and installation methods. The anticipated diameter of the jet-grout columns was between 2.4-3m. The diameter of the deep soil mix columns was between 0.9-1.8m. All columns were designed to begin at elevation 980 and to terminate at elevation 1016.
The purpose of the test section program was to provide the United States Army Corps of Engineers (USACE) with information for use in evaluating the effectiveness of jet-grouting operations to reduce the seismic risk associated with the project. Additional objectives included selecting the material proportions and installation methods to be used during construction of the cut-off wall and to identify geophysical methods capable of measuring the thickness of the main cut-off wall (not constructed) after installation.
The downstream test section site was located approximately 198m south of the existing dam. Tuttle Creek Lake and dam lie in the Kansas River Valley. The deposits found within the downstream test section consist of alluvium from the Big Blue river and surrounding tributaries.
Geophysical borehole installation
Treviicos South, in cooperation with USACE, selected a single jet-grout column (JG26), a single deep soil mix column (DSM23), and an area in the native soil outside of the slurry cut-off wall around the downstream test section for geophysical testing.
Geotechnology selected a layout for the survey arrays that involved a series of five boreholes at both of the selected modified soil columns and four boreholes in the native soil area. The boreholes were aligned in a square pattern centered upon the column that was to be tested. The array was oriented such that the lines connecting the boreholes at the perimeter of the array would be either perpendicular or parallel to the axis of the modified soil column rows. The perimeter boreholes were located 0.6m from the design boundary of the modified soil columns. The borehole locations were staked by a licensed surveyor supplied by Treviicos South. A layout of the test section and borehole arrays appears in Figure 1.
The boreholes were drilled with CME 55HT and CME 750 drill rigs utilising a combination of six and a quarter inch hollow stem augers and six inch mud rotary drilling techniques to advance the borings to the desired depths. The boreholes were installed to an approximate depth of 15m. The boreholes were cased using four inch schedule 40 flush thread PVC casing and grouted with a Portland cement, bentonite, and water grout mixture to bond the casing to the borehole wall. The casings were capped and grouted from the bottom up using a one inch diameter tremmie pipe inserted between the casing and boring wall. Deviation logging was performed on all boreholes on the day following installation. Cross-hole seismic tomography and cross-hole seismic velocity surveys were conducted during periods of inactivity at the site to avoid vibrations created by construction equipment.
Borehole deviation logging
Borehole deviation measurements were conducted using a Mount Sopris MGXII series portable digital logging system with a 2DVA-1000 deviation probe. Two data sets were recorded for each of the 14 boreholes. The first data set for each borehole was designated as the down log, starting from the ground surface and progressing down to the bottom of the casing. The second data set for each borehole was recorded from the bottom of the borehole up to the surface. These logs were designated as the up logs.
The results of the borehole deviation measurements were used to determine the distance between boreholes at depth, which in turn facilitated the reduction and interpretation of cross-hole seismic velocity survey data and the interpretation of the cross-hole seismic tomography data.
Cross-hole seismic tomography
The cross-hole seismic tomography method involves generating Compressional P-wave seismic energy at several depth intervals within each borehole and measuring the seismic wave travel-time at multiple geophones situated along the length of an opposing borehole. By moving the source and keeping receiver positions constant, a dense cluster of seismic raypaths is surveyed. The raypaths are interpreted by analysing the differences in elapsed travel-time from the source to the geophones. A velocity model of the volume covered by the seismic raypaths can then be constructed and interpreted using tomographic inversion.
Data acquisition and processing
Cross-hole seismic tomography surveys were conducted using a 24-channel Geometrics, SmartSeis 24 seismograph. Data was collected on the full length of each tube-pair (ten tube-pair permutations for the DSM column, ten for the jet-grout column, and six for the native soil array). A string of twelve hydrophones spaced 1m apart was placed at the bottom of the receiver borehole. A mechanical borehole hammer was then placed in the source borehole and activated at each 1.5m interval along the entire length of the borehole. Approximately ten successive signals were made with the borehole hammer at each depth interval. The seismograph recorded and stacked each of the repeated geophone responses.
The hydrophone string was raised to the top of the borehole and the signal was repeated at each source depth. Approximately 200 viable raypaths per tube-pair were collected during the tomographic survey. There were a total of approximately 2000 viable raypaths for each of JG26 and DSM23, and approximately 1200 raypaths for the native soil area. Data were processed using TomTime and GeoTomCG seismic and tomography software to pick the P-wave arrival times. Next, Geotechnology developed a velocity model and generated a topographic inversion. A three-dimensional data volume was created and representative two-dimensional planes through each data volume were imaged using Slicer Dicer. A collection of the two-dimensional planes within the three-dimensional data volume and their positions relative to the applicable boreholes appears on Figures 2-4.
Summary of results
The calculated P-wave velocities in the deep soil mix array indicate a general increase of velocity with depth in the upper 4m of the columns surveyed. An increased velocity ‘silhouette’ is located in the center and towards borehole 3 from 4m to the bottom of the survey volume (16.5m). As illustrated in Figure 2, the calculated velocity increases to 1494m/sec in the center of the volume and toward borehole 3. A triangular-shaped anomaly exhibiting greatly reduced velocity in one corner of the bottom of the survey volume is representative of an absence of data due to a reduced depth of one of the boreholes and is not indicative of actual velocity data.
In the jet-grout array the calculated velocities appear to increase at different locations within the volume of the survey. The tomography models appear to indicate that a 20% increase in velocity was identified at the center of the survey volume, beginning 0.3m below the surface grade and passing to the bottom elevation of the survey (15m). A velocity increase appears as a ‘silhouette’ on the planes between tube pairs 1-4, 3-4, 1-3, 2-4, and in the horizontal cross-sections in Figure 3. An increase in velocity in the direction of boreholes 3 and 4 begins at 9m below grade and extends to the bottom of the survey area as well.
As illustrated in Figure 4, the calculated P-wave velocity appears to increase with depth from 1371.6m to 1706.9m/sec in the native soil area. The velocities are uniformly distributed over the entire horizontal cross-section of the survey array and did not indicate any localised velocity changes across the survey volume at any selected depth. A triangular-shaped anomaly exhibiting greatly reduced velocity at the bottom of the survey volume is representative of an absence of data due to a reduced depth of one of the boreholes and is not indicative of actual velocity data.
The cross-hole seismic data indicated a general increase in seismic velocity with depth in the native soil, and a divergence in velocities between the native soil and the modified soil columns. This information may be used by USACE in determining the seismic properties of the native soil and column materials.
The expected P-wave velocity of the modified soil columns was approximately 2133.6 to 2438.4m/sec. The calculated P-wave velocity through the jet-grout column (762m/sec) and the deep soil mix column (1493.5m/sec) were well below the expected range. The variations in calculated P-wave velocities with depth, location, and test method seem to indicate that the shallow water table may alter the signal travel times when using seismic methods. The ability of the cross-hole tomography to image the columns suggests that the relative difference between the seismic velocity of the modified soil and the native soil can be used to resolve features, but the accuracy of the actual calculated P-wave velocities is questionable.
In addition, the cross-hole seismic tomography data indicated the presence of a higher velocity feature within the center of the survey volumes for the deep soil mix and jet-grout column locations. No localised velocity changes were identified in the horizontal direction of the native soil survey volume. The identified features began to appear in the data at depths that corresponded with the design elevation of the tops of the columns (0.3m below grade). The column locations and dimensions were verified by direct measurement of the excavated columns at a later date as illustrated in the main photograph. These factors would appear to indicate that the cross-hole seismic velocity surveys were able to identify the presence of the columns within the survey volumes. However, the resolution of the identified features was insufficient to determine the length or diameter of the columns with accuracy. The resolution might be improved with the use of a higher frequency energy source. The shorter wavelengths produced by this source would increase the likelihood of resolving smaller features and more precisely identify the column boundaries. The reliability of data at the edges of the tomograms is also limited due to the reduction in the ray path density at the edges of the survey area.
Justin Kindt, PE. Project Manager and Casey Jones, PE, RG, Branch Manager, Geotechnology, INC. 11816 Lackland Rd., Suite 150, St. Louis, MO 63146. Tel: +1 314.997.7440. www.geotechnology.com