Deep mixing application in New Orleans13 April 2007
Mark Lee Woodward presents details on three deep mixing applications for Task Force Guardian, the unit assembled by the US Army Corps of Engineers to repair 354km of the Greater New Orleans Federal hurricane and flood protection system
Hurricanes Katrina and Rita caused severe damage to the Hurricane and Flood Protection system within Orleans, St. Bernard and Plaquemines Parish, US. The damage included overtopping, breaches and washouts of both walls and earthen levees. Task Force Guardian of the US Army Corps of Engineers (USACE) was assembled to repair approximately 220 miles (354km) of Greater New Orleans Federal hurricane and flood protection system damages resulting from Hurricane Katrina to pre-storm conditions by 1 June 2006.
A total of 59 construction projects have been awarded, with deep mixing being used on three of these. Two of these projects entail using deep mixing to decrease lateral active earth pressures and increase lateral passive earth pressures at closure structures under construction at the mouths of interior drainage canals in New Orleans, Louisiana. The other deep mixing application is being used beneath an earthen hurricane/river flooding protection levee in Plaquemines Parish to improve the overall foundation competency with respect to landside slope stability.
The deep mixing method involves the blending of a binder such as lime, cement, slag, fly ash, etc. into the soil through a hollow stem auger and mixing tool arrangement to produce round 'columns' of treated soil. Deep mixing methods are usually broken into two methods depending on how the binder is introduced into the soil. The dry mix method uses a mixing tool which is rotated downward into the soil at high speed while compressed air is blown through the binder port in the tool shearing the soil. Once required depth for ground improvement is reached, the direction of the tool is reversed and dry binder is pneumatically blown into soil as the mixing tool is withdrawn. Moisture is drawn from the in-situ soil and hydration of the binder begins followed by a longer term pozzolanic reaction. In the wet mix method, the binder is premixed with water at a ratio between 0.6 to 1.5 by weight of water to weight of binder to create slurry which is pumped into soil under relatively low pressures. The wet method normally produces columns of higher strength compared to dry mixed columns. The wet method rigs uses multiple augers, which allows for excellent overlap compared to the single shaft rigs widely used in dry mixing. The wet method produces significant spoils compared to a relative absence of spoils with the dry mix method. The dry mix method may be used in areas where water supply is unavailable or is of dubious quality, although it is common practice to construct wet mixed columns in marine applications using seawater.
Applications for the deep mixing method include stability and support, seepage cutoff, and seismic retrofit. All of the deep mixing applications for Task Force Guardian fall into the stability and support category, although USACE has used the wet deep mixing for seepage cutoff in California in the levees along the American river and for seismic retrofit in South Carolina to mitigate liquefaction in alluvial fill adjacent to the Clemson dam.
Prior to Hurricane Katrina, the New Orleans District had conducted a successful deep mixing test section1,2 using dry mixed columns adjacent to the Inner Harbor Navigation Canal just east of downtown New Orleans in 2003. This test section simulated the loading of a full earthen embankment on two different replacement ratio column configurations. With the knowledge and experience garnered during construction of the test section columns, test section loading, test section evaluation and research into deep mixing, it was a logical and comfortable step to implement deep mixing for Task Force Guardian projects. Once need for ground improvement was analysed and strength and layout of columns designed, Drs. Donald Bruce and Massaaki Terashi were placed under contract to perform an internal technical review of the deep mixing designs and construction.
Outfall canal interim closure structures
Deep Mixing is being utilised at two of the three outfall canal interim gated flood control structures, which are under construction at the mouths of the 17th Street, Orleans, and London Avenue Outfall Canals. These structures will prevent a storm surge in Lake Pontchartrain from entering the outfall canals while allowing drainage.
17TH Street outfall canal
The 17th Street Interim Closure Structure (Project OEB 09) is located at the mouth of the 17th Street Canal.
It was determined by the designers that the in-situ soil underlying an area 50ft (15.2m) wide on both the flood side and protected side of the closure structure would have to be improved from an average shear strength of 280 psf and a low of 100 psf to 1000 psf to alleviate active pressure on the structure and increase passive resistance behind the structure. Ground improvement was deemed necessary to an Elevation of -65.0, where the Pleistocene horizon is encountered. Ground improvement was also determined to be required along the east side of the canal to increase stability of fill to be placed against the floodwall. A schematic of the structure is shown in Figure 1, and Figure 2 shows areas to be deep mixed.
Plans and specifications denoted the area to be treated in plan, column configuration array; specified depth of treatment, specified required strength and continuity of columns, and requirement that quality control testing is performed on 3% of the production columns. The specifications were geared towards dry mixing, but left the door open for any ground improvement method. The strength required in the contract was set at 120-psi unconfined compressive strength. Assuming a 33% replacement ratio and applying a factor of safety of 2.0 to account for columns being less than 28 days old when placed into service and a safety factor of 1.5 to account for material properties; the required unconfined compressive of the columns was determined as follows: ((3 3 2000 psf) /144 in2/ft2) 3 2.0 3 1.5 = 120 psi. Jet grouting was specified between the structure and deep mixing. The layout of the columns specified in the plans is shown in Figure 3. It included continuous overlap (mass stabilisation) for the first three column rows parallel to the sheetpile, followed by a grid pattern with a replacement ratio of approximately 33%.
The contract for the 17th Street Closure Structure was awarded to Boh Brothers Construction who subcontracted the Ground Improvement work to Hayward Baker Incorporated. Hayward Baker proposed to use dry mix deep mixing using a single axis, LCM dry mixing rig and crawler mounted tank system from a barge. They embarked on their test programme by mixing samples from three different types of soil from seven borings taken in the area with three different binder loadings and two different binder mixtures. The binder mixtures used were a blast furnace slag/cement mixture and a pure cement mixture. The foundation to be treated was broken down into three distinct soil layers: an upper organic layer with a thickness of 5 to 7ft (1.5-2.1m); a middle Intermediate clay layer with a bottom Elevation of -45.0; and a bottom layer to the bottom of the treated zone. Figure 4 shows test results of 3-day, 6 or 7-day and 14-day breaks with projections to 28 days, and Figure 5 shows actual 28-day and 56-day strengths. In the upper organic layer, the 28th day strength projection and actual strength was significantly below the strength requirement. Hayward Baker proposed to double mix this zone, effectively doubling the binder dosage by repenetrating and injecting binder. Additional bench testing was performed on the upper organic layer with a mixture of 75/25-slag/cement ratio by weight and binder dosages of 350kg/m3 and 400kg/m3.
Results shown in Figure 6 indicate that the required strength criteria could be met by double mixing this layer. The only binder dosage that met the strength criteria in the middle layer was 175kg/m3 using 100% cement. In the bottom layer all mixes/dosages met the required 28-day strength.
Two quality control methods were proposed consisting of coring of the columns using a triple tube sample with a coarse diamond bit and PORT. PORT (pull out resistance testing) or RCPT (reverse column penetration testing) is a method where a vane attached to a cable is left at the bottom of a column during installation and subsequently pulled up through the column at a later time. The shear strength of the column can be determined by dividing the measured resistance force by the area of the top of the vane and a bearing factor.
Twelve test columns were installed just outside of the treatment zone with the following makeup:
• Four Columns with 200kg/m3 single treatment of upper organics, three with PORT vanes left 1m below column and one with cable only.
• Four Columns with 200kg/m3 double treatment of upper organics, three with PORT vanes left 1m below column and one with cable only.
• Four Columns with 175kg/m3 double treatment of upper organics, three with PORT vanes left 1m below column and one with cable only.
The vanes were left a meter below the column and the vanes were pulled up to the bottom of the columns at t = 3 days to debond the cable from the column. Cable only columns were installed so that the cable friction could be evaluated and subtracted from the total resistance during pulling of the vanes.
PORT testing was attempted at t = 7 days, but the cable failed before the vane could be pulled through the column. The maximum viable strength for the cable is 85 psi. The decision was made to use insitu core retrieval as the sole measure of column strength. The test columns were cored at 16, 18 and 20 days providing 32 samples for UCS testing at 34 days. The unconfined compression strengths are presented in Figure 7.
Production columns were installed using a binder load of 200kg/m3 and double mixing the upper organic layer. A layout of the deep mixed columns (2.6ft [0.8m] diameter), jet grouting (5ft [1.5m] diameter) is shown in Figure 8.
Three percent of the columns were cored between 16 and 20 days for testing at 28 days.
Orleans canal interim closure structure
The Orleans Avenue Interim closure structure is located at the mouth of the Orleans Avenue Canal (Figure 9).
It was determined by the designers that the in-situ soil underlying the protected and flood side of the closure structure would need to be improved. As with the 17th Canal Closure Structure, Ground Improvement was specified to increase passive resistance to provide greater lateral support and to reduce active pressure on the structure. Ground Improvement was specified to an Elevation of -50.0. The foundation is characterised by a canal bottom of organic clays underlain by a hydraulic fill layer comprised mainly of silt to Elevation -20.0, followed by Lacustrine clay to -35.0, underlain by a Beach sand to Elevation -45.0, under laid by Bay Sound clays to below bottom of ground improvement.
The specifications were very similar to those for the 17th Street Canal Closure Structure. It contains 377 single columns on the protected side and 262 single columns on the flood side. Again jet grouting was specified between the Deep Mixing and the structure. Schematics of the structure are shown in Figures 10, 11, and 12.
The construction of the interim closure structure was awarded to Gilbert Southern, who in turn hired Raito to perform the deep mixing using the wet mixing method. Raito used one triple axis CDSM machine type 608 on a barge and a land based semi-automatic grout batching plant. Raito used a grab soil sample taken from the upper 5ft (1.5m) below the mud line to conduct a laboratory bench scale test. This material was considered to be the worst material within the foundation to El. -50 and this needing the most binder to raise the insitu strength to 120 psi UCS. The soil sample was mixed cement to form a grout with a water content ratio of unity by weight with the following binder loads: 250kg/m3, 300kg/m3 and 350kg/m3. Only the 3 day and 7 day breaks were available before Raito started deep mixing operations, due to the urgency of construction. The values of the 7 day breaks were 75 psi for both the 300 and 350kg/m3 loadings, Raito decided to used a 350kg/m3 cement loading with a water cement ratio of 0.8 for the production columns to insure that required column strength would be met, there was no time in the schedule for inferior columns. The results of the laboratory testing with 28 day strengths are shown in Figure 13.
Raito created a deep mixing plan for both the flood side and the protected side. The plan for the protected side closely matched contract drawings, however, due to ease of access and time of construction Raito's proposed plan for the flood side was different enough from the contract drawings, that a review by USACE and the designers (URS) was required. It was determined that Raito's proposed flood side plan satisfied the intent of the Deep Mixing and was approved. Raito proposed methods of quality control were wet grab samples and coring. Wet Grab samples were taken daily from a box affixed to the central axis with mechanical capability to open and close at desired depth. Initially grab samples were taken at mid-depth; this was adjusted to the upper ten feet to insure that the weaker layer be tested. Material from the grab sample box was pored into three inch diameter molds six inches in length. Test results all met the contract strength requirement. Results are presented in Figure 14.
Raito cored 3% of the elements or four elements per side of the structure using its 2-1/2 tube core barrel.
Results are presented in Figure 15. The 28 day strengths far exceeded the contract strength requirements. Photos of cores are presented as Figure 16.
Plaquemine Parish Homeplace Levee Setback
An existing Mississippi river floodwall, located southwest of Port Sulphur, Louisiana, approximately 65 miles (105km) down river from New Orleans, was damaged during Hurricane Katrina and will be demolished and a full earthen levee will be constructed behind it, further from the Mississippi river. Because of low landside slope stability factors of safety, less than unity, deep mixing was selected to improve the overall foundation competency to avoid a large landside berm, necessitating road relocation and taking of private property. The road relocation and additional right of way would affect a 1900ft (579m) reach.
Only three borings existed within the area to receive ground improvement, so three additional 5-inch diameter undisturbed borings were drilled in January 2006. With regards to the geologic profile, the surface and shallow subsurface is generally composed of natural levee, marsh and swamp, interdistributary, intradelta, and prodelta deposits. The entire area is overlain by approximately 7-10ft (2.1-3m) of natural levee deposits, which extend down to approximately -8ft (-2.4m) NGVD. Natural levee deposits generally consist of oxidised fat and lean clays and silts with relatively low water contents and higher compressive strengths than the surrounding environments. Approximately 5ft (1.5m) of interdistributary deposits underlie natural levee deposits in boring PSV-12UT. Interdistributary deposits consist of fat clays with lenses and layers of lean clay, silt, and silty sand. Marsh and swamp deposits underlie natural levee and interdistributary deposits. Marsh and swamp deposits are approximately 5ft (1.5m) thick and extend down to approximately -15ft (4.6m) NGVD. These deposits consist of very soft to medium fat clays with lenses and layers of lean and organic clay, and wood. Marsh and swamp deposits are underlain by interdistributary deposits which are approximately 28-42ft (8.5-12.8m) thick and extend down to approximately -53ft (-16m) NGVD. Swamp deposits are found within interdistributary deposits in boring PSV-10UT from approximately -28 to -34 ft (-8.5m to -10.4m) NGVD. Swamp deposits averaging 4ft (1.2m) thick are also found from approximate station 418+00 to the end of the study area between -28 to -38ft (-8.5m to -10.4m) NGVD. Underlying interdistributary deposits are intradelta deposits which are up to 20ft (6m) thick. These deposits range in elevation from -38ft to -60ft (-11.6m to -18.2m) NGVD and consist of silt, sand, and silty sand with lenses and layers of clay. Intradelta deposits are underlain by prodelta deposits, which extend to the bottom of the borings. Prodelta deposits consist of medium fat clay with lenses and layers of lean clay.
In order to raise the landside stability factor of safety above 1.30 using the method of planes, the foundation soil had be raised from an average value of 270 psf to 2300 psf. In order to achieve 2300 psf, the design called for replacing 30% of the soil beneath a 35ft (10.7m) width beneath the levee which equates to 800mm column panels at 7ft (2.1m) center to center spacing. The columns need average shear strength of 50 psi or 7200 psf, as indicated in
csv = cns (1 - Ac ) + clc Ac (1)
where csv is the shear strength of the total soil volume, c is the shear strength of the natural soil, Ac is the part of the total shear surface covered by the columns, and clc is the shear strength of the columns
For a 30% replacement ratio,
2,300 psf = 270 psf (1 - 0.30) + clc (0.30)
clc = 7,040 psf = 49 psi
Each panel row contained 17 columns at 25.5 inch spacing providing 6 inches of overlap between columns. At 7ft (2.1m) spacing between column panels, the 1900ft (579m) reach required 272 panels for a total of 4624 individual columns.
The advertised levee enlargement job covered 4.5 miles (7.2km) of levee enlargement and one mile of levee setback that included the 1900ft (579m) of deep mixing. The job was awarded to the Shaw Group, who subcontracted with Hayward Baker to install and test the ground improvement. Hayward Baker opted to use the dry mix method to install the columns. It started the mix design by taking nine borings at 200ft (60m) intervals along the area to be deep mixed. Material from these borings was used for bench scale testing; the samples were homogenised as follows:
• A Soil - approx depth - 0 to 11ft (0 to 3.4m)
• B Soil - approx depth - 11 to 23ft (3.4 to 7m)
• C Soil - approx depth - 24 to 30ft (7.3 to 9.1m)
• D Soil - approx depth - 30ft (9.1m) to design depth
The A Soil was mixed with 175kg/m3 slag/cement binder loading to make eight samples for testing at 5, 7, 28 and 56 days. The A soil was also mixed with 175kg/m3 slag/cement binder and 10% additional moisture by weight of soil to make 8 samples for testing at 3, 7, 28 and 56 days. The B soil was mixed with 175kg/m3 slag/cement binder and 200kg/m3 slag/cement binder to make a total of 16 samples for testing at 3, 7, 28 and 56 days. The same was performed with the C soil.
Extensive mixing energy and/or effect of heat testing were performed on the D soil. The D soil was mixed with 150kg/m3, 175kg/m3 and 200kg/m3 slag/cement binder to make a total of 36 samples for testing at 3 or 4, 7, 28 and 56 days. All of the nine mixes mentioned above were prepared using a Hobart mixer on medium speed (setting 2) for one minute of mixing time. Twenty four additional samples were prepared with the D soil using 175kg/m3 slag/cement binder however one set of 12 samples were prepared using a Hobart mixer on low speed (setting 1) for 45 seconds of mixing time and another 12 samples were prepared using a Hobart mixer on medium speed (setting 2) for 10 minutes of mixing time. Test results are shown in Figures 17 and 18.
Nine test columns were initially installed at the downstream end of the ground improvement area, each with a 175kg/m3 slag/cement binder loading. PORT vanes where also installed in each column, PORT testing was attempted on each initial test column. The vane could be pulled through the column at four hours and six hours after installation, however at eight hours and beyond PORT testing was not successful. Port Test results for the three successful PORT tests are presented in Figures 19, 20, and 21. Two of the initial test columns were cored.
The specifications require that one column out of every other panel be tested at 3ft (0.91m) intervals along the length of the column. This quality control requirement is being accomplished by coring and unconfined compression testing.
Deep mixing has proven to be a viable method to effectively improve the competency of soils in Southeast Louisiana. Both the dry and wet deep mixing methods have demonstrated that they can be used to substantially raise the insitu shear strength of the soil several orders of magnitude. All three of the deep mixing applications for Task Force Guardian required rapid construction techniques, construction sequencing, and logistical problems associated with working in confined work areas. Both deep mixing installers proved willing to meet these challenges by rapidly mobilising equipment and resources to the metro New Orleans area, quickly embarking on exploration and testing programs, and insuring quality deep mixed columns by using higher than normal binder loadings.
Figure 1: Location and vicinity map of 17th Street Canal Interim Closure Structure Figure 1 Figure 2: Plan view showing location of 17th Street Canal Interim Closure Structure Figure 2 Figure 3: Schematic of 17th Street Canal Interim Closure Structure Figure 3 Figure 4: Plan view of 17th Street Canal Interim Closure Structure. Deep mixing is shown in light blue Figure 4 Figure 5: Layout of columns specified in contract specifications Figure 5 Figure 6: 17th Street Bench scale 3, 7 and 14 day testing results with projection to 28 days Figure 6 Figure 7: 17th Street Canal bench scale test results Figure 7 Figure 8: 17th Street Canal, double mixed bench scale test results of upper material Figure 8 Figure 9: Photo of PORT vane Figure 9 Figure 10: 17th Street Canal - 34 day UCS - test columns Figure 10 Figure 11: Deep mixing at 17th Street Canal shown in light blue, jet grouting shown in red Figure 11 Figure 12: Ground Improvement at 17th Street Canal, light blue circles indicate 2.6' diameter dry mixed columns, and 5' diameter jet grout columns are indicated in red Figure 12 Figure 13a: Core samples from the 17th Street production columns Figure 13a Figure 13b: Core samples from the 17th Street production columns Figure 13b Figure 14: Location and vicinity map of Orleans Canal interim closure structure Figure 14 Figure 15: Plan view showing location of Orleans Canal interim closure structure Figure 15 Figure 16: Orleans Canal deep mixing Figure 16 Figure 17: Orleans Canal closure structure Figure 17 Figure 18: Orleans Canal Closure struture Figure 18 Figure 19: Orleans Canal closure structure Figure 19 Figure 20: Results of Orleans Canal Bench scle testing Figure 20 Figure 21: Orleans Canal floodside deep mixing Figure 21 Figure 22: Orleans Avenue protected side deep mixing Figure 22 Figure 23: Orleans Canal Wet Grab sampling box Figure 23 Figure 24: Wet Grab sampling UCS results Figure 24 Figure 25: Orleans Canal coring UCS results Figure 25 Figure 26: Orleans canal cores Figure 26 Figure 27: Location map showing Homeplace levee enlargement Figure 27 Figure 28: Homeplace geologic profile Figure 28 Figure 29: Section view of Homeplace deep mixing Figure 29 Figure 30: Homeplace bench scale test Figure 30 Figure 31: Homeplace bench scale test results Figure 31 Figure 32: PORT test result - Homeplace (unreduced to account for cable friction) Figure 32 Figure 33: PORT test results, Homeplace - four hours Figure 33 Figure 34: PORT test results, Homeplace - six hours Figure 34 Author Info:
Mark Lee Woodward, PE, Lead Geotechnical Engineer, CEMVN-ED-F, P.O. Box 60267, New Orleans, LA 70160-0267. Email: [email protected]