Read Cement Deep Soil Mixing Remediation of Sunset North Basin Dam text version


Rebecca F. Barron1, Christopher Kramer2, W. Andrew Herlache2, Jon Wright1, Howard Fung3, and Chu Liu3


Cement Deep Soil Mixing (CDSM) is a state-of-the-art soil improvement method developed in Japan and gaining acceptance in the United States. The design team selected CDSM for the rehabilitation of the supporting embankment to San Francisco's largest potable water reservoir, Sunset Reservoir. This project marks the first time the California Department of Water Resources, Division of Safety of Dams has approved CDSM to remediate a potentially weak foundation and thereby improve the seismic stability of an earth embankment dam. The design for the seismic remediation was completed in November 2004. Construction began in June 2005 and was completed June 2006. Sunset Reservoir is located in a densely populated residential neighborhood of San Francisco, with homes and businesses located immediately downstream of the dam. The reservoir has a storage capacity of approximately 177 million gallons or 543 acre-feet and supplies roughly 60 percent of the City's water. The San Francisco Public Utilities Commission (SFPUC) owns and operates the reservoir. The seismic stability of the dam, a critical City facility, was the focus of the project. The SFPUC's criterion is for the dam and reservoir to remain functional after a major earthquake on the San Andreas Fault, which is located 5 kilometers from the site. The 74-foot high earth embankment was built in 1938 on top of saturated native sands and silts. Geotechnical investigations revealed that these soils could be susceptible to significant strength loss during a major earthquake. This paper describes the design and construction process, including the initial geotechnical investigation, the selection of CDSM as the preferred remediation technique from various alternatives, the design of the remediation scheme, and construction issues and lessons learned. The paper also addresses the collaboration between the different agencies, consultants, and specialty contractors that led to the successful completion of the project.



California Department of Water Resources, Division of Safety of Dams Fugro West, Inc. (formerly with Olivia Chen Consultants, Inc.). 3 San Francisco Public Utilities Commission


This paper describes the evaluation, selection, design, and construction of Cement Deep Soil Mixing (CDSM) to retrofit the northwest embankment dam of Sunset North Basin in San Francisco, California. Originally developed in Japan, CDSM is a technique for in-situ ground improvement that is gaining acceptance in the United States. CDSM is a process to improve soil by injecting grout through augers that mix it with the soil, forming in-place soilcement columns. Figure 1 presents two photographs of the CDSM equipment in operation on the embankment.

Figure 1. CDSM Rig at Sunset Reservoir Sunset Reservoir is a lined and covered offstream reservoir located in the southwestern region of San Francisco. It is the largest potable water reservoir in San Francisco, supplying roughly 60 percent of the City's water supply; it is owned and operated by the San Francisco Public Utilities Commission (SFPUC). The reservoir comprises the North and South Basins, which have a total storage capacity of approximately 177 million gallons or 543 acre-feet. The embankment, located at the northwest corner of the north basin, is 74 feet high. The north basin was built in 1938 using a combination of cut and fill. The south basin was built in 1960 by excavating into relatively stable bedrock of the Franciscan Complex. Sunset Reservoir's storage capacity and dam height place it under the jurisdiction of the California Department of Water Resources, Division of Safety of Dams (DSOD). In 1998, as part of the reservoir seismic retrofit, SFPUC initiated a field investigation and engineering study to evaluate the seismic performance of the dam and reservoir. In 2000, the SFPUC's Consultants and DSOD concluded that a maximum earthquake on the San Andreas or Hayward Fault could result in strength loss of the foundation soils below the northwest embankment of the North Basin.

Roles of Various Parties

Owner The San Francisco Public Utilities Commission is a department of the City and County of San Francisco that provides water, wastewater, and municipal power services to San Francisco. Under contractual agreement with 28 wholesale water agencies, the SFPUC also supplies water to 1.6 million additional customers within three Bay Area counties. As owner and operator, the SFPUC retained the geotechnical and civil engineering consultants, managed the design phase, and provided contractual oversight during construction. It was also responsible for monitoring the Quality Assurance and Control for the project and working with the specialty contractor to minimize the disruption to the neighboring residents and business, while maintaining the safety and the availability of water to its customers.

Geotechnical and Civil Engineering Consultants Olivia Chen Consultants, Inc. (Consultants) was retained by the SFPUC to provide geotechnical and civil engineering services for the project. The Consultants performed the geotechnical investigation, conducted engineering analysis, and designed the remediation scheme. The Consultants and the SFPUC jointly developed the project construction plans and specifications. The Consultants also provided geotechnical observation and consultation during construction. Specialty Contractor The construction contract was awarded to Gordon N. Ball/Yerba Buena Joint Venture (GB/YB), the general contractor. The CDSM portion of the project was subcontracted to Raito, Inc., a construction company specializing in subsurface in-situ soil improvements using CDSM and Dry Jet Mixing methods. Based in Japan, Raito, Inc. has been one of the leading developers of CDSM technology. GB/YB conducted all excavation safety monitoring, conducted all earthwork activities, and managed all interaction with the owner, consultants, and regulatory agencies. Raito, Inc. was responsible for conducting the CDSM ground improvement and all verification testing. Regulatory Agency The Division of Safety of Dams is a division of the California Department of Water Resources responsible for the supervision of dams larger than a certain size. It is DSOD's responsibility to assure that these dams and reservoirs are designed, constructed, operated, and maintained safe from failure due to natural static forces, as well as hazards such as floods, earthquakes, landslides, destructive seepage, and serious operational malfunctions. To accomplish this, DSOD maintains a staff of engineers, geologists, and seismic specialists who perform independent engineering analysis and monitor the construction of jurisdictional dams. DSOD staff also performs periodic safety evaluations of all jurisdictional dams to monitor their maintenance and performance. Currently, there are approximately 1250 dams under the State of California's jurisdiction. DSOD worked with the SFPUC and the Consultants, providing input for the geotechnical investigation and reviewing the field and laboratory results. DSOD performed an independent engineering analysis of the existing embankment and foundation, and the designed remediation. DSOD reviewed and approved the remediation design and monitored the construction work with respect to dam safety. Continuing changes in the fields of seismology and geotechnical engineering, as well as breakthroughs in modeling technology, have resulted in a greater understanding of how dams behave under both static and seismic conditions. DSOD is continually working to maintain a program that utilizes the latest state-of-the-art technology.

Original Reservoir Construction

Sunset North Basin was built in 1938 utilizing cut and fill techniques, resulting in a storage capacity of 275 acre-feet and a 74-foot high earthfill embankment along its north and west sides. The embankment has a crest length of 2,300 feet and a crest width of 11 feet. The upstream (interior) slope is 3:1 (horizontal to vertical); the downstream slope is 2.5:1, with a 10-foot wide bench at mid-height. The dam has 2.0 feet of freeboard. The embankment was founded on a thick layer of dune sand, which was loose in some areas. According to as-built drawings and construction photos, the topsoil was stripped within

embankment's footprint. Where the natural ground was below the bottom of the proposed reservoir, sandy material excavated from within the basin's limits was used as backfill. The embankment materials came from the north and south basin excavations and from the construction of Balboa Reservoir. The embankment materials consist of clayey gravel with sand (GC) and clayey sand with gravel (SC) with a rock fill shell on the downstream slope. Based on the specifications and construction photos, the finer grained material was placed by dump trucks, and then spread into 6-inch horizontal lifts before being compacted by dozers with tandem sheepsfoot rollers. The rock fill was dumped in 3-foot layers and then compacted with power rollers. Any voids between the separated rock fill were filled in with materials excavated from the reservoir.

Geotechnical Investigation

The geotechnical investigation of Sunset North Basin Dam included a review of the previously performed work, field investigation, and laboratory testing. The Consultants performed the field investigation with input from the SFPUC and DSOD. Under the direction of the Consultants, an independent geotechnical laboratory performed geotechnical testing. Previous Investigation Because of the Baldwin Hill Failure in 1963, Division 3 of the Water Code was amended to include both new and existing offstream dams. A field investigation and embankment stability analysis were required to bring Sunset North Basin Dam into State jurisdiction. An initial investigation and analysis were performed in 1969 with six rotary wash borings drilled to depths up to 90 feet. Standard Penetration Testing (SPT) was performed. Uncorrected blow counts ranged from 11 to refusal, with 5 percent below 20. Soil samples were taken using SPT and pitcher barrel samplers. Laboratory testing was performed on remolded and pitcher barrel samples to determine material strength properties and to characterize the embankment and foundation materials. The analysis found the embankment to have adequate safety factors for both static and pseudo-static models. However, portions of the foundation were identified as possibly being susceptible to strength loss during a large seismic event. Eight piezometers were installed in four of the borings to monitor the phreatic surface of the embankment and foundation. Recent Field Investigations In 1998, as part of the SFPUC's reservoir seismic retrofit, DSOD requested an updated stability analysis of the embankment. A phased field investigation was begun in 1998 with seven borings drilled up to 41 feet deep. SPT testing was performed. Uncorrected blow counts ranged from 4 to refusal, with 41 percent below 20 for the foundation soils. Soil samples were taken using the SPT and Modified California samplers. Laboratory tests were performed to obtain soil properties and material strength data. Subsequent phases were performed in 2000, 2002, and 2002-2003 to obtain additional blow count data and material properties, and to better delineate the lithology of the foundation materials. A total of 20 additional borings were drilled, ranging in depth from 21 and 95 feet. The materials encountered ranged from the rock fill of the embankment, to poorly graded and silty sands to the formational bedrock. The approximate locations of the borings from the investigations are presented on the Boring Location Plan, Figure 2.

Figure 2. Boring Location Plan

Geology, Subsurface Conditions, and Seismicity

Regional Geology Sunset Reservoir is located on the San Francisco Peninsula between the San Andreas Fault, 5 km to the west, and Hayward Fault, 25 km to the east. The peninsula is underlain by isolated Jura-Cretaceous outcrops of Franciscan Complex surrounded by Quaternary age Colma Formation (sand, silt, and gravel) and younger dune sand. The Franciscan Complex is a structural unit. Site Geology Bedrock at the Sunset North Basin embankment site consists of sheared sandstone and shale of the Franciscan Complex (mélange), which is overlain by surficial deposits of silty sand with clayey sand lenses, ranging in thickness from 5 to 15 feet. This silty sand is thought to be derived from the Pleistocene Colma Formation. Dune sand (poorly graded fine sand) overlies the silty sand layer and ranges in thickness from 10 to 30 feet. The depth to bedrock ranges from 85 feet from the crest of the dam to 33 feet from the toe of the dam.

Soil Conditions Nine lithologic units were delineated during subsurface investigation of the northwest corner of North Basin. These units are described in the following paragraphs in order of decreasing elevation, and are shown in Figure 3 below.

Figure 3. Cross Section A Silty Sand 1 (Fill): Approximately 5 feet of topsoil consisting of medium dense to dense silty sand with gravel. Gravel 1 (Rock Fill): Five to 25 feet of gravel and rock fill below the downstream face of the embankment. The rock fill consists of medium dense to dense gravel, with varying amounts of clay, sand, and rock fragments up to about 3 feet in diameter. The source of this material appears to be select Franciscan Complex material excavated from the south basin area. Gravel 2 (Fill): Forty-five to 50 feet of medium dense clayey gravel with some sand within the main embankment. This unit also contains rock fragments up to 3 feet in diameter. The source of the material appears to be select Franciscan Complex material excavated from the south basin area. Poorly Graded Sand 1: A 10- to 15-foot layer of loose clean fine-grained native dune sand with localized medium dense areas was encountered below the downstream toe of the embankment. Poorly Graded Sand 2: A 5- to 20-foot thick layer of dense to very dense clean fine-grained dune sand, generally located below the crest the embankment. Poorly Graded Sand 3 (FILL): Poorly graded dune sand placed and compacted during the original construction to fill in low areas of the foundation. Silty Sand 2: A 3 to 7-foot thick layer of loose to medium dense saturated silty sand located below the poorly graded dune. Silty Sand 3: A 3 to 11 feet thick layer of medium dense to dense saturated silty sand, with interbedded zones of clayey sand was encountered below Silty Sand 2 layer. Franciscan Complex: The Franciscan complex underlying the silty sand is composed of moderately to severely weathered, closely fractured to crushed, moderately strong to friable, and moderately hard to friable interbedded sandstone and shale.

Groundwater Prior to construction of the CDSM, there were 10 open standpipe piezometers measuring ground water levels along the northwest corner of the embankment, with sets of two to three piezometers nested together in some locations. The SFPUC performed regular measurements of groundwater levels in the piezometers. Historical piezometer readings have shown that the regional groundwater levels within the foundation vary little over time. Historical readings also indicate the possibility of minor localized seepage (or perched water) within the embankment. Seismicity and Faulting The major active faults in the San Francisco Bay Area are the San Andreas, Hayward, San Gregorio, Rodger's Creek, and Calaveras. No active or potentially active faults are known to cross the site. The distances from the site to major active faults in the area and their anticipated maximum magnitude are presented in the following table. Approximate Distance from Site (km) 5 10 25 40 37 42 Maximum Moment Magnitude 8.0 7.3 7.1 7.0 6.8 6.8

Fault San Andreas San Gregorio Hayward Rodger's Creek Point Reyes Calaveras

These active faults are capable of generating strong ground shaking at the site during an earthquake. A moment magnitude (Mw) 8 earthquake on the San Andreas Fault is the controlling seismic event. Using 84th percentile deterministic attenuation relationships, the Consultants calculated a bedrock peak ground acceleration (PGA) of 0.97 gravities (g) at the project site. DSOD accepted the seismic parameters after performing an independent evaluation with comparable results.

Seismic Design

Sunset Reservoir is a critical facility for the City of San Francisco. The embankment performance criteria were for the reservoir to be able to sustain minor damage, remain operational, and suffer no catastrophic release of water during or after the earthquake. The project team selected a maximum permanent seismic deformation criterion of 6 inches, to be calculated using Newmark sliding block methodology. The Consultants developed site-specific horizontal response spectra, peak ground acceleration, and spectrally matched input acceleration time histories for use with the Newmark model. Some cracking of the reservoir liner is anticipated at this level of deformation, but the reservoir should be operable.

Pre-Treatment Conditions

Loss of Strength The soils of concern for Sunset North Basin Dam were two silty sand layers, Silty Sand 2 and Silty Sand 3, which are located below the water table. The Consultants evaluated these layers for potential significant strength loss using the Youd et al. (2001) method. Input parameters included a moment-magnitude 8.0 earthquake on the San Andreas Fault and a peak ground acceleration of about 0.97g at the site. The results of the evaluation are presented in Figure 4. The Consultants concluded that Silty Sand 2 was susceptible to significant strength loss, and that the majority of Silty Sand 3 was not, except in some localized areas. The predicted consequences of significant strength loss (prior to the retrofit) included settlement and slope instability. Using Ishihara's 1993 method, the Consultants estimated that earthquake-induced settlement could be up to about ½ foot. This settlement would not be uniform, and could reduce the reservoir freeboard in some locations. The settlement could also result in localized cracking of the liner and some distress to the roof structure.

Figure 4. Corrected N Values for Silty Sand In addition to settlement, the temporary reduction in shear strength associated with significant strength loss could result in embankment slope instability during or after a significant earthquake, if mitigation measures are not implemented. For their independent evaluation, DSOD selected a more conservative approach. They selected a higher phreatic surface, extending from maximum reservoir elevation to the toe of the dam, to model a case of an ineffective reservoir liner. In the DSOD model, the lower portions of Poorly Graded Sand 1 are also considered to be susceptible to significant strength

loss. The following table provides a range of uncorrected field blow counts, averaged 30th percentile (N1)60 blow counts, and average percent fines content for the three layers. Averaged (N1)60 (30th percentile) Percent Fines 18 2 8 29 24 44

Soil Layer NField Poorly Graded Sand 1 4 to 55 (20 avg) Silty Sand 2 4 to 44 (14 avg) Silty Sand 3 12 to 73 (37 avg)

Seismic Stability The Consultants evaluated the stability of the northwest embankment under static and seismic loading conditions, including considering a reduction in shear strength of the Silty Sand 2 during seismic loading. Idealized Subsurface Cross Section A, depicted in Figure 3, was the critical cross section because the critical layer is at its deepest and the embankment is at its tallest. They used computer software and the modified Bishop's method of slices to evaluate a large number of circular and wedge-shaped failure surfaces. Published data (Youd, 2003) indicate that significant strength loss of soils will require some time to develop after the beginning of an earthquake. In addition, looser soils will lose strength sooner than denser soils. The Consultants divided Silty Sand 2 into two sublayers: looser and denser. Stage 1 represents the condition prior to and just after the beginning of the seismic event; at this time strength loss has not yet occurred in any layer. Stage 2 represents the condition after the looser sublayers lose strength. Stage 3 represents the condition after the denser sublayers lose strength. Typical failure surfaces for the cases analyzed are shown in Figure 5.

Figure 5. Critical Pre-Treatment Failure Surfaces The following table presents the strength parameters used to analyze the three stages of seismic induced strength loss in the Silty Sand 2. Looser Sublayers = 30 degrees Su,r = 210 psf Su,r = 210 psf Denser Sublayers = 30 degrees = 30 degrees Su,r = 450 psf

Analysis Stage 1 Stage 2 Stage 3

The parameter Su,r represents the seismically-induced undrained residual shear strength, which the Consultants estimated using the correlation presented in the paper by Seed and Harder (1990). First, the Consultants evaluated stability for each stage, assuming reductions in shear strength during Stages 2 and 3, but no seismic inertia loading. They then used acceleration time histories in combination with the results of slope stability analyses to estimate the magnitude of permanent deformation using a sliding block model (suggested by Newmark, 1965). The critical factors of safety and yield accelerations are presented in the following table. The most critical failure surfaces are shown on Figure 5. Factor of Safety Critical Failure Surfaces (no seismic inertia loading) Stage 1, Downstream 2.6 Stage 2, Downstream 1.9 Stage 3, Downstream Stage 3, Upstream 0.6 4.7

Yield Acceleration (g) 0.46 0.26 (Not applicable when factor of safety is less than 1.0) 0.63

For downstream failure surfaces (away from the reservoir), the consultant concluded that severe deformations were possible as a result of the earthquake. For upstream failure surfaces, the consultant estimated that permanent deformations would be less than about an inch. The Consultants further concluded that embankment foundation improvement would be required to improve the seismic performance of the downstream slope of the embankment. DSOD performed an independent analysis. Using the Consultants' geotechnical data, they performed a post-seismic static analysis on circular and wedge-shaped failure surfaces. Residual strengths were assigned to the potentially weak layers based on the Seed and Harder method (1990). The post-seismic factor of safety was found to be less than one, confirming the risk of significant deformation after a strong seismic event.

Selection of Preferred Remediation Method

The Consultants initially considered the following remediation alternatives: Cement Deep Soil Mixing (CDSM), jet grouting, compaction grouting, vibro-replacement, permeation grouting, and excavation and recompaction. CDSM was selected after eliminating the other options based on the following considerations: · Jet grouting, was eliminated because the Consultants concluded that the process could not be controlled to an acceptable degree and because preliminary estimates indicated it would be more costly than CDSM. · Compaction grouting is known to have limited effectiveness in soils with relatively high fines content. The Consultants concluded that compaction grouting was not appropriate because the target layer, Silty Sand 2, had a median fines content between 28 and 35 percent. · Vibro-replacement also is less effective for soils with high fines contents. In addition, the Consultants concluded that vibro-replacement equipment might have difficulties penetrating embankment soils to effectively treat the target layer. · Permeation grouting was eliminated because Silty Sand 2 was not permeable enough to allow uniform penetration of the grout.

The SFPUC considered excavation and recompaction of the entire embankment to be prohibitively expensive and lengthy compared to other methods and felt it could not afford to have the reservoir taken out of service for an extended period of time. Considering the limitations of other methods, the SFPUC accepted the Consultants' recommendation to use CDSM for foundation remediation. CDSM is a process whereby soil is improved by injecting grout through one or more augers that simultaneously mix the soil, forming in-place soil-cement columns as shown in Figure 7.


Figure 6. CDSM Construction Procedure The CDSM would need to be performed in a regular grid of in-place columns to effectively improve the target soils. Level benches would have to be cut into the embankment in order to operate the CDSM rig. The grid would have to be designed to provide subsurface drainage paths to minimize buildup of groundwater behind the soil cement columns. Advantages of CDSM include: · High strengths can be achieved in the final soil cement product. · CDSM has been used on similar projects to address seismic stability issues associated with significant strength loss, such as the Jackson Lake Dam and the Clemson Upper and Lower Diversion Dams. It has also been used on several projects within the Bay Area, such as at the Port of Oakland. · The zone of improvement can be controlled more effectively than jet grouting. · Proven confirmation testing methods exist for assuring quality.


The process would have a low impact on reservoir operations.

Disadvantages include: · Temporary construction benches would have to be cut into the embankment during installation. · The process generates spoils that must be used onsite or hauled offsite.

Final Design

Consultants Approach The Consultants performed a series of stability evaluations to develop a CDSM scheme that would meet the static and seismic performance objectives. They evaluated multiple layouts with varying material properties prior to selecting a suitable treatment pattern. The resulting treatment pattern is depicted in plan view in Figure 7, and in profile in Figure 8. The layout consists of multiple 47-foot square grids (blocks) of CDSM columns that are placed in three rows (treatment zones) aligned parallel to the longitudinal axis of the embankment. The blocks consist of exterior and interior walls formed by adjacent CDSM columns, as shown on Figure 7.

Figure 7. Treatment Plan The configuration required that the CDSM have an average 90-day unconfined compressive strength of at least 400 pounds per square inch (psi), and that 25 percent of Silty Sand 2 be improved by CDSM. Available laboratory data indicated that a 90-day strength of 400 psi will result if the 28-day strength is at least 300 psi. The resulting weighted-average shear strength of the improved layer is approximately 7200 pounds per square foot (psf), or about 75 percent higher than the strength of the underlying Silty Sand 3. The Consultants recommended the discrete block layout, with gaps between the individual blocks to provide a conduit for regional groundwater. The recommended treatment zones extend at least 5 feet below the bottom of Silty Sand 2, or to the top of bedrock when shallower. Site geometry allows three rows of CDSM blocks near the center of the embankment, decreasing to two rows on the east and west ends. Where only two rows were possible, the CDSM is keyed into the underlying bedrock to provide additional resistance to sliding.

Figure 8. Post-Treatment Failure Surfaces (Section A) In developing the design described above, the Consultants revised the slope stability model, assigning improved soil properties for zones to be treated using CDSM. They ignored the strength of unmixed soil within and between the grid blocks. Regarding the potential for CDSM cracking, the Consultants concluded that: the structural integrity of the blocks would limit cracking associated with ground shaking; potential failure surfaces crossing through the blocks would experience insignificant deformation; and failure surfaces passing below the blocks, which bypass the CDSM, would have acceptable levels of deformation. The Consultants repeated the staged analysis approach with the revised model. Two typical failure surfaces are shown on Figure 8. In the downstream direction, the lowest static factor of safety prior to earthquake loading (Stage 1) was 2.8, indicating the embankment is stable for this case. After Stage 3, the minimum factor of safety was still 2.8. For this case, the minimum yield acceleration is 0.41, resulting in less than 6 inches of calculated deformation after the controlling seismic event. Based on the analysis, the Consultants concluded that the CDSM treatment scheme meets the seismic performance objectives.

DSOD Approach DSOD's analysis conservatively adopted a saturated foundation and a higher phreatic line within the embankment, taking into consideration the potential for the water table to rise due to the proposed remediation. The critical circular failure surface was determined by using the Modified Bishop method, while the critical block-slide failure surface was determined by the Spencer method. Post-seismic factors of safety for circular and wedge-shaped failure surfaces were found to be greater than 2, with a corresponding minimum yield acceleration of 0.35. The results of the stability analysis indicated acceptable factors of safety for the postseismic condition and relatively high yield accelerations. DSOD estimated the seismic deformation of the embankment with the proposed foundation improvement using the Makdisi and Seed simplified procedure (1978). Based on the calculated yield accelerations and the estimated maximum horizontal acceleration along the critical failure surface, DSOD estimated the seismic deformation to be less than 1 foot. This is within the current freeboard of 2.0 feet, satisfying the embankment performance criterion of not allowing a catastrophic release of water during or after the earthquake. Therefore, DSOD agreed with the proposed design provided the CDSM was extended upward into the Poorly Graded Sand 2 layer.

Construction and Monitoring

Construction Overview Construction of the embankment stabilization began in May 2005 and was substantially completed by December 2006. The work consisted of site demolition and grubbing, embankment excavation, CDSM construction, embankment reconstruction, restoration of site improvements, and re-landscaping. Embankment excavation, CDSM construction, and embankment reconstruction were completed for each treatment zone prior to beginning the next. The construction proceeded from Treatment Zone 3 to Treatment Zone 1, from the bottom of the embankment to the top. The CDSM equipment required a 60-foot wide, level surface to operate, so the contractor had to excavate a flat bench for each of the treatment zones. Safe excavation and recompaction of the soil was a major logistical challenge. Figure 9 depicts the temporary construction of Bench 2. Material from the excavation was used to backfill Bench 3. The following constraints impacted the construction activities: space limitations for onsite temporary stockpiling of excavated embankment soil and construction staging, the requirement that the reservoir be kept in service during construction, and construction within a densely populated residential neighborhood. These issues were managed through close coordination between GB/YB, Raito, Inc., SFPUC construction and engineering staff, DSOD engineers, and the Consultants. During construction, detailed monitoring revealed some issues that necessitated modification of construction procedures and monitoring activities. Of particular interest are the issues associated with the embankment excavation and CDSM construction. Temporary Embankment Excavation To address the stability issues associated with the temporary embankment excavation, a detailed monitoring program and emergency response plan were developed. The monitoring program included daily monitoring of 10 survey hubs set along the crest of the embankment, weekly monitoring of three inclinometers installed adjacent to selected survey hubs, weekly

monitoring of existing piezometers, and daily reconnaissance of the excavation to look for any indications of possible instability. Planning for emergency response included defining the roles and responsibilities of each party, disseminating emergency contact information for key individuals, discussing possible scenarios for heightened concern, establishing monitoring thresholds, developing communication protocols for addressing concerns, and preparing for possible remedial measures that could be implemented if stability concerns arose. During construction, rapid dissemination and evaluation of the monitoring data was achieved using email and flexible communication protocols established by the SFPUC. Deformation monitoring generally revealed small movements (less than ¼ inch) and the temporary cut slopes remained stable. On occasion, the survey data suggested more significant movements, but these were not consistent with reconnaissance observations and inclinometer data and were later determined to be errors in surveying and/or data entry. Localized seepage zones were observed on all three temporary benches; however, piezometer monitoring revealed no significant increase in the phreatic surface within the embankment. Drains were installed along each of the benches during reconstruction to ensure the embankment remains dry.

Figure 9. Temporary Construction Bench 2 Cement Deep Soil Mixing The challenges during CDSM construction included limitations on space for construction staging and rig operations, coordination during excavation and embankment reconstruction, assessing the effectiveness of the soil improvement, and evaluating the depth of penetration into weathered bedrock. The first two challenges required close coordination between GB/YB and Raito, Inc., and were met without significant cost or schedule impacts to the project. The latter two challenges were of particular concern to the SFPUC, the Consultants, and DSOD. To assess the effectiveness the CDSM improvement, a detailed monitoring and testing program was established. The program included test sections utilizing several different mix designs with varied cement contents and water-cement ratios, monitoring of various parameters during CDSM construction, full depth coring of the completed CDSM columns, unconfined compressive testing on selected CDSM core samples, test borings prior to CDSM construction or adjacent to completed CDSM columns, and test redrilling of cured CDSM.

The monitoring program generally indicated that the CDSM at the site satisfied the requirements in the project plans and specifications. Specifically, the following points were noted: · The completed CDSM satisfied the geometric requirements for plan location, configurations, depth, inclination, and overlap of adjacent elements. · Coring and testing indicated the minimum 28-day unconfined compressive strength of the CDSM within the treatment zones was 179 psi (minimum 120 psi required). The average strength for any complete core, in the treatment zone, was found to vary between 322 and 1177 psi (minimum 300 psi required). · Core recovery and uniformity of the CDSM was found to be satisfactory. The recovery averaged 96 to 100 percent within the treatment zone (minimum 85 percent required) and the percent of unmixed soil ranged from 0 to 15 percent (maximum 15 percent allowed). While the CDSM achieved the goals established for the ground improvement, evaluating the penetration of the CDSM into the bedrock was of particular concern and required additional testing. The specifications required that some of the CDSM elements penetrate at least 5 feet into bedrock; other elements were only required to reach the top of the bedrock. During the design phase, the project team anticipated that various CDSM drilling parameters, correlated to boring logs, could be used to assess the depth of penetration into bedrock. However, during construction, the project team could not establish a suitable correlation. In addition, comparison of boring logs and practical drilling refusal suggested that the CDSM rig was not able to consistently penetrate into bedrock. Therefore, additional testing and analyses were conducted. The depth to bedrock below the site was known to vary significantly over a short distance. The original boring locations were not surveyed; therefore, the exact depth to bedrock at a given CDSM element location was uncertain. Additional borings were drilled adjacent to CDSM elements to determine the depth to bedrock and to allow correlation with the CDSM drilling parameters. The drilling parameters included the actual vertical penetration rate (speed), drive motor amperage, energy index (amperage divided by penetration rate), the measured load on the cables that supported the augers and motors, and the visual observation of slack in the cables once the load was being supported by the subsurface materials. The following table summarizes the data where the new borings were drilled. Depth of Penetration into Bedrock (feet)* Large Drop in Spike in Zero Penetration Energy Cable Cable Rate Index Load Slack 1.6 1.6 0.6 1.5 4.2 4.2 4.2 3.9 -0.5 -0.5 -2.5 -2.6 3.5 3.5 3.5 3.5

Boring No. CB-1 CB-2 CB-3 CB-4

CDSM Element No. Z3-B6-T14 Z3-B3-T11 Z3-B2-T13 Z3-B1-T13

Actual Depth of Element into Bedrock (feet) 3.3 4.9 -0.5 3.5

* The depth below the bedrock surface, based on a significant change in the specific drilling parameter.

From the data, it was apparent that a suitable correlation between the depth to bedrock and drilling parameters could not be reliably established. In addition, the actual depth of penetration into bedrock at the boring locations varied significantly. The authors believe that

these observations were likely the result of the variation in the depth of the bedrock along the width of one CDSM element (13 feet) and variations in the hardness and degree of weathering of the bedrock. Based on these findings, a test was performed to assess how strong the bedrock had to be before encountering practical drilling refusal. The test involved re-drilling a cured CDSM element while monitoring the drilling parameters. The test revealed that the CDSM rig was capable of drilling into material with an unconfined compressive strength of at least 700 psi. Additional seismic deformation analyses confirmed that acceptable seismic performance would be achieved if the CDSM columns were founded on rock with an unconfined compressive strength of at least 700 psi. On the basis of the borings and test results, the termination criteria was modified from requiring 5 feet of penetration into bedrock to achieving cable slack and a penetration rate of less than or equal to 0.1 feet per minute. CDSM Costs Raito Inc. estimated the CDSM work would be competed in about 4.5 months for a cost of about $3.7 million. Actual construction of the CDSM was completed over a period of six months, with 4.4 months for the CDSM construction and 1.6 months of down time. 21.4 million pounds of cement were used to construct 4353 soil cement columns, for a total cost of about $4.5 million.

Conclusions and Lessons Learned

Cement Deep Soil Mixing was selected as the preferred alternative to mitigate a potentially weak foundation. This was a new technology for the SFPUC and DSOD; therefore, it was important to develop clear and effective specifications for its construction. It was also important for there to be enough flexibility in the procedures to address variations in the conditions encountered during construction. Thorough mixing, material strength, and termination criteria were considered key components required to achieve a quality product. Verification testing was performed to ensure that the requirements for these criteria were met. Another essential element to the success of the project was the close coordination between the SFPUC, the Consultants, the Contractors, and DSOD staff. Key design considerations when applying newer technologies include: · Designers must understand detailed construction aspects of the new technology in order to anticipate the full range of potential problems related to specific site conditions. · The unique details associated with newer technologies require special attention to identify all of the potential failure modes. · A thorough geotechnical investigation, performed in phases, is critical to understanding the subsurface site geology and to obtaining reliable material properties and strength data. · Early communication and coordination between the consultant and the regulatory agency is essential to promote understanding of the technology and regulatory requirements. The project was a success from both the design and construction point of view. The site is located in a densely populated neighborhood in San Francisco, and the project was

completed with minimal disruption to the residents. The design of the CDSM treatment to improve the weak foundation was able to meet the seismic performance objectives that were established for the project. The CDSM construction was completed in a manner that was consistent with the design, with strengths that were greater than required. The embankment excavation and reconstruction were successfully completed on a very constricted site, while accommodating the CDSM construction schedule. And finally, the changed conditions that were encountered during construction, and the challenges they presented, were evaluated and successfully addressed by the SFPUC, the Consultants, and DSOD. The project was also successful for other reasons as well. A new technology was successfully used for a site that would have been difficult to remediate otherwise. The CDSM technology proved successful in addressing the difficult site constraints while achieving the required improvement to the foundation soils. This project presents a model for taking newer technologies from conceptual ideas to effective implementation through sound engineering and construction practices. The SFPUC, the Consultants, the Contractors, and DSOD, all worked together with an understanding of each other's needs and concerns to achieve the common goal of a good final product. Lessons learned during construction include: · Redundant/confirmatory instrumentation monitoring is necessary to help identify the cause(s) of unanticipated conditions and to assess their implications regarding slope stability. · Emergency response planning is essential to identify the roles and responsibility of the various parties involved, to disseminate emergency contact information, to define communication protocols that allow for rapid evaluation and decision making, and to anticipate and prepare for possible problems that develop during construction. · Prompt communication between the contractor, SFPUC representatives, engineering team, and regulatory agencies are necessary to identify and respond to potential problems in a timely manner. · CDSM can be an effective tool to mitigate stability concerns associated with seismically induced significant strength loss for embankment dams. · Construction monitoring, coring, and strength testing can be used to verify that the required soil improvement has been achieved. · CDSM rig penetration into weathered rock can vary significantly and drilling parameters may not reliably indicate the depth of penetration into rock. · The width of the CDSM elements, the variation in the surface of the bedrock, and the type and degree of weathering of the rock need to be considered when establishing CDSM depth termination criteria. · A CDSM depth termination criterion that is based on the performance of the CDSM rig may be necessary to achieve the desired results without imposing excessive requirements on the CDSM subcontractor. However, the criterion must be calibrated to the particular equipment being used and the subsurface conditions encountered. · The use of proper CDSM construction equipment and methods can result in significant strength increases and relatively uniform ground improvement in loose to medium dense sands.


The authors wish to thank Dr. Robert Pyke for providing input and guidance during the design and Dr. David Yang of Raito, Inc. for information provided during design and construction.


1. Makdisi, F.I. and Seed, H.B., (1978) Simplified Procedure for Estimating Dam and Embankment Earthquake-Induced Deformations, ASCE Journal of the Geotechnical Engineering Division, Vol 104, No. 7, pp. 849-867 2. Newmark, N., (1965) Effects of Earthquakes on Dams and Embankments, Geotechnique, Vol. 15, No. 2. 3. Olivia Chen Consultants, (2001) Sunset Reservoir North Basin Embankment/Foundation Stability Assessment. 4. Seed, R.B. and L.F. Harder. (1990) SPT-Based Analysis of Cyclic Pore Pressure Generation and Undrained Residual Strength. J.M. Duncan, Editor, Proc. H. Bolton Seed Memorial Symp., Vol. 2, 351-376. 5. Tagasoft, TNMN Software, (2003). 6. Tokimatsu, K. and Seed, H.B., (1987) Evaluation of Settlements in Sands Due to Earthquake Shaking, Journal of Geotechnical Engineering Division, ASCE, Vol. 113, GT8. 7. Youd, T. L., (October 23, 2003) Influence of Increased Pore Pressure and Liquefaction on Spectra Acceleration, H. Bolton Seed Memorial Lecture, Foster City, California (to be published in ASCE's Journal of Geotechnical Engineering). 8. Youd, T.L., Idriss, I.M. Andrus, R.D. Arango, I., Castro, G., Christian, J.T., Dobry, R., Liam Finn, W.D.L., Harder, L.F., Jr., Hynes, M.E., Ishihara, K., Koester, J.P., Liao, S.S.C., Marcuson, W.F., III, Martin, G.R., Mitchell, J.K., Moriwaki, Y., Power, M.S., Robertson, P.K., Seed, R.B., Stokoe, K.H., II, (2001) Liquefaction Resistance of Soils: Summary Report from the 1996 NCEER and 1998 NCEER/NSF Workshops on Evaluation of Liquefaction Resistance of Soils, ASCE, Journal of Geotechnical and Geoenvironmental Engineering, V. 127, No. 10, p 817-833 9. Bishop, A.W., (1955). "The Use of the Slip Circle in the Stability Analysis of Slopes." Geotechnique, Vol. 5, No. 1, March, pp. 7-17. 10. Ishihara, K. (1993). "Liquefaction and flow failure during earthquakes." Geotechnique, 43(3), 351-415 11. Spencer, E., (1967). "A Method of Analysis of the Stability of Embankments Assuming Parallel Inter-Slice Forces." Geotechinque, Vol. 17, No. 1, March, pp. 11-26.


Cement Deep Soil Mixing Remediation of Sunset North Basin Dam

19 pages

Report File (DMCA)

Our content is added by our users. We aim to remove reported files within 1 working day. Please use this link to notify us:

Report this file as copyright or inappropriate


You might also be interested in

Aberg, B
Cement Deep Soil Mixing Remediation of Sunset North Basin Dam