Abatement of soil liquefaction under existing structures

H. R. AI-AlusiPressure Grout Company, Hayward, California, U.S.A.

To be presented at:IS Tokyo '96~~~~~--=~~~"C'"'~~ ••.

Second International Conference onGround Improvement Geosystems- Grouting and Deep Mixing -

May 1996

ABSTRACT: Liquefaction mitigation under a functioning structure represents a challenge to the-engineer andthe geotechnical contractor and accentuates cost, time, and disturbance of the facility's use. Acces~is the mostimportant aspect of such a project. Two case histories where soil liquefaction was required to-be abated under

",,, existing structures are presented. . ".~._. In Case One, a filter building was underlain by a 3.7 meter liquefiable sand layer. The approach taken to

reach the target soil was to drill horizontally about 28 meters, and extrude compaction grout in l.S-meter stagesto densify the soil.

In Case Two, a liquefiable sand layer and a silty/clay fill layer were identified under an existing multi-storybuilding. The. building was supported on old wood piles. Two systems were selected for ...this project;permeation grouting for sands and lense grouting for silty/clay layer.

In conclusion, it was possible, using cost-effective methods, to mitigate the risk of liquefaction under existingstructures without interrupting the use of the facilities.

1 I:t-..TTRODUCTION

Improvement of liquefiable soils follows a numberof methods and techniques that are well establishedin the industry. These available methods andtechniques become very limited and restricted forlocations where a structure is in place but underlainby liquefiable soils. In a situation where a structuremust continue to function, the restrictions multiply.Liquefaction mitigation of a soil layer under afunctioning facility represents a challenge to theengineer and the geotechnical constructor.

Almost every project for abatement is unique in itsapproach. Consultation and close coordinationbetween the owner of the facility, the geotechnicalengineer, and the specialty contractor are oftantamount importance for these projects.

This paper presents two case histories wheregrouting methods were successfully used underexisting structures for the abatement of soilliquefaction in a cost-effective way withoutinterfering with the operation of the facilities.

2 CASE ONE: COMPACTION GROUTING FORSAND DENSIFICA TION

Compaction grouting has been successfully used forsand densification for liquefaction mitigation(Mitchell & Wentz 1991). The process involves thecontrolled injection of a stiff sand/cement mixture tovolumetrically displace the soils to increase itsdensity. Grout stiffness is related to the slumpvalue, ASTM C143-78. For a controllable groutingprocess, this va1ue should be' maintained to less than5 centimeters (Al-Alusi 1994). The success of thisprocess is hinged on having the ability to maintainthe grou t near the point of injection byvolumetrically displacing the soils without fracturingthem. Theoretically, compaction grouting loses itssignificance upon fracturing the soils. Groundsurface monitoring for vertical and/or horizontaldisplacements (in cases involving slopes) is a mustfor evelY application.

.. ."".

~HDUSANOITIOAKS

Fig. 1: Site Location

2.1 Site conditions and liquefaction potential

An exploration to ascertain the soil conditions forexpanding a wastewater reclamation plant in LosAngeles, Califomia (see Figure 1), identified aliquefiable soil layer approximately 3.7 meters thick.The expansion program included adding more filtersto the existing filter building, which was 25 X 30meters in plan. Although the new filters could befounded on a deep engineered fill, the liquefiablesoils under the existing filters building needed to beapproached differently.

PROPOSED -f<~"~ '-"FILTERS

meters'49.5

PLAN

.....l53,,,TypicalPROPOSED

FILTERSGroundSurface EXISTING FILTERS

137

"Silty Sa~d.Sandy Silt. . .•.,,'," '" ,1!riySd:~,~~~~L-..,....,.,-.....:B::::o=tlO7m~of~Ga:::;.::II::::er;~=,..,.L.1

;~~ ~~;~~:.S;:.t:;~1~;:1f1:~.~;.~~f-=:;'''3 Bottom of Filters

Sand to Silty Sand

I : •• I Potentially Liquefiable Soils- - _ ? Interpreted Geological Contact

W Soils to be improved by Compaction Grouting

~ 50ilsto be removed and Recompacted with5tJucwraiFillSECTION A-A [Not to Scale)

Fig. 2: Plan and Section of Areas Designated to beDensified using Compaction Grouting

Bordering the structure on three sides were otherfacilities in such close proximity that made itimpossible to excavate for a horizontal drillingoperation. The remaining side of the structure thatwas available for horizontal drilling was a shortside. An excavation of about 10 meters was madewith conventional soldier beams and lagging. Theground water level was dropped using a traditionalpump and sump method.

The project site was located in a seismically activearea of Southern California. The dominating fault,located approximately 4.8 kilometers miles from thesite, was considered to be capable of generating amagnitude 7.5 earthquake. Probabilistic seismic riskanalyses estimated the Peak Ground Acceleration at0_35g and 0.6g for a design life of 100 years. Thedensification process became critical due to thebuilding's sensitivity to settlement (Harding Lawson1990).

Soil borings and Cone Penetrometer Test (CPT)soundings encountered a 2.5-meter layer ofrelatively loose sand. The geotechnical engineers'site seismicity evaluation and cyclic shear strengthstudy indicated that this layer was potentiallyliquefiable when subjected to the groundaccelerations of design-earthquakes, see section inFigure 2.

Seismically-induced settlements were found to belikely to occur in the loose to medium dense layer(3.7 meters thick). These settlements werecalculated to be on the order of 5 to 13 centimetersusing methods proposed by Tokimatsu and Seed(1987).

2.2 Approach, Drilling, and Grouting

The project specifications called for extending thedensification process for a 1.5-meter strip aroundthree sides of the building, see Figure 2. For thisstrip a single row of vertical injections, spaced at 1.5meters, was used. The installation of theseinjections served to confine the grout under thebuilding and to provide correlations between testresults and the amount of the injected grout (grouttake) under the building for quality assurance. Thesame termination criteria that were used forhorizontal grouting, as discussed below, were used.

Two rows of horizontal injections were installed1.8 meters apaTt, and centralized in the middle ofthe liquefiable zone, see Figure 3. The spacingbetween injections was about two meters. Eachinjection was extended horizontally for the full

ExistingFiltersBldg.

(Not to Scale)

Fig. 3: On-Site Details During Compaction Grout Operationand Section through the area to be compacted

length of the existing filters, reaching a maximum of26 meters. A partial drilling of 3.5 centimeter holewas first made with the aid of drilling foam. Uponcompletion of the hole a nominal five centimeterclose-ended pipe was driven in.

Compaction grout extrusions started at the far endof the excavation side. A staging of 1.5 meters wasused. The sequence of grouting was by theprimary-secondary injection method, alternatinginjections between the two rows.

At each stage, grout was injected until one of thefollowing criteria was met:

1. A maximum pressure of SSOOkPa is reached atthe point.

2. Inception of ground or structural uplift.3. A quantity of 400 liters is pumped at a given

stage.Throughout this operation the high pressure

criterion controlled the process. Grout takes rangedbetween 11 and 37 liters. A total of 19 horizontalinjections were completed.

Throughout the grouting operation, horizontal andvertical, a laser survey system was employed tomonitor any uplift.in the structure or ground surface.None was detected at any stage.

2.3 Test results and discussions

Standard Penetration Tests were utilized to evaluatethe effectiveness of the grout densification processaround the perimeter of the existing building. Acomparison of test results conducted before andafter the densification process is shown on Figure 4.Since the SPT results were erratic, a correlationbetween the SPT increase and the calculated densityincrease caused by the added grout mass in a givenvolume of soil was made using publisliedcorrelations (Holeyrnan & Wallays 1984,Winterkorn and Fang 1991, Bowles 1982). Thesecalculations revealed that the relative density of thesoil was raised from about 45 to 70 percent, andthat the SPT values were raised from a range of 10to 20 blows to a range of 30 to 40 blows. These

3.05

I ------After lBefore

--...~ --- ---1----- l/ \'- \ I

l \

; t~\

\ \

<uV V (r", \t-------.

~

/'

~ r--r---. <:............•• ••..•.......•..

~ -r-,---.1-. '- "-..• -

(J)LOJ~OJEc 6.10LwD-O)o

8.1515 30 45 60 75 90

Test No.115 30 45 60 75 90 15 30 45 60 75 90

2 SPT 315 30 45 60 75 90

4

Fig. 4: Standard Penetration Test Results for After and Before Compaction Grouting

25 50 75Depth in 0mere~O r------------------------------.

o Boring HB-1II Boring HB-2+ Boring HB-3

Boring HB-4* Boring HB-5-:. Boring HB-6

3.05;~y LAssumed ...••.

\ • *+'\ ' ;....,..~~~~~~~"atero

\+( -----\-r\ --;>,Criticoi Penetrotion\ .". \7 Resistance\ .* \. II;W~~\

.!.\• I

..•• if.!.····1:·~\ ..II·I .:, ....;.\.• :........~.

\ EARTHQUAK~ EVENTSI~ M,bS-' .: '::Ir.;,,';j\ .. (O.35g) ·t:\/lO.Sg)

6.10

9.15

Standard Penetration Blows

Fig. 5: Soil Improvement Relative toEarthquake Events

results indicated that the targeted soils wereimproved to well above the critical penetrationvalues required for a magnitude 6 earthquake event,and at or near that of a magnitude 7 earthquakeevent, see Figure 5.

It was not possible to perform blow count tests orany sounding tests, such as static cone penetrations,under the structure. Although other sounding tests,such as ultrasound and cross-hole geophysical tests,were available, it was concluded that such methodswould not reveal sufficient useful results, especiallyin zones where known gravels existed. Theeffectiveness of the grouting program in this zone

Fig. 6: Plan showing Chemical and Lense Grouting Zones

was evaluated by comparing the amount of groutinjected in various stages, which was calculated tobe about three percent of the total volume of soil, tothe theoretical density improvement.

3 CASE TWO: CHEMICAL AND LENSEGROUTING AROUND EXISTING PILES

Liquefiable soils, due to their particle sizedistribution and resulting high permeability, areusually amenable to chemical/permeation grouting.Chemical grouting, simply put, is a pure permeationgrouting, which utilizes two or more materialcomponents whose chemical reaction results in ahardened matrix within the soil mass. In saturatedsoil masses, the grout is expected to displace waterfrom the soil pores. In partially saturated soils itdisplaces air and water. Confinement and' control ofthe grout are two key elements of a successfulchemical grouting job. They can be achieved byworking in a designed pattern, with or without aprescribed gel time (i.e., time required after mixingto start hardening),

Lense grouting is a soil fracturing technique wherea cement slurry grout is injected at an initial highpressure of 700 to 3000 kPa, then reduced until apredetermined amount of grout is injected. Inman-made fills and alluvial deposits, near horizontalfractures are achieved using engineered tips at thebottom of the injection pipe to facilitate fractureinitiation (Al-Alusi 1994).

3.1 Building and Foundation Conditions

A seismic upgrade program was to be implementedfor an eight-story concrete and masonry structuremeasuring 25 x 25 meters at the basement level.The building .was located on a corner in thedowntown area of San Francisco, California. It waserected in 1907 in the area tha.t had experiencedground failure during the April 18, 1906 earthquake.During the Lorna Prieta earthquake of October 17,1989, the building sustained structural damage. Thefoundation consisted of pile groups, as shown onFigure 6. Timber piles, 46 centimeters in diameterand 10+ meters long, are believed to have supportedthis structure. The pile caps and floor were ofreinforced concrete construction. Below thebasement bottom, the soils were predominantlysandy silt with clay and some rubble fill (layer 1),for a distance of 2.5 meters, then a loose gravelly

• <

o Street Level

-28m Basement Level( (

Sandy Silt PileCap yW.L.WithClay .~

Lense-5.5m Grouting

Gravelly TimberSand +- Piles->- Chemical

Grouting

-8.3mClay

N I[ ot to Sea e)

. Fig. 7: Typical Section

sand layer which extended another 2.8 meters (layer2), where a clay layer started, see Figure 7. Theblow count (standard penetration test) was between1 and 6 for layer 1, and between 2 and 13 for layer2. Based on the very low blow count and highground water level at which was at about thebasement level, it was determined that liquefactionwas most likely to occur in these sand formationsduring an earthquake comparable to the design eventof magnitude 7.

3.2 Restrictions and criteria of treatment methods

Compaction grouting could eliminate the potentialfor liquefaction of the soils below the basementfloor slab. However, because of the presence of thetimber piles and the likelihood that some of themwere in a partially deteriorated condition, coupledwith the high pressures inherently associated withcompaction grouting, this method was deemed to beunacceptable even though. it would.. have been themost cost effective.

In order not to affect the timber piles by highgrouting pressures, a chemical system was selectedthat would solidify the sands of layer 2 with a lowstrength grout to render the material non-liquefiable.A sodium silicate based grout was used with a geltime of ten minutes and an ultimate unconfinedcompressive strength of the grouted sand of 190 kPawas selected and installed. Because layer 1 was notsusceptible to chemical grouting, this method waslimited to layer 2 only. The exact location of thepiles within each group (cap) was unknown. Fewattempts were made to define the pile locations,

which revealed that the actual locations were not asshown on the available plans. Consequently, thedesign of the chemical grout system was made toconfine each group in an isolated cell, then injectmore grout within the confined cell to refusal, seefigure 6. Refusal is defined as a grout pressure of1300 kPa or a predetermined amount of grout basedon the theoretical volume of voids within the cellusing a porosity of 0.35.

The procedure followed was that of a closed endpipe vertically jetted into the ground using water.At the proper depth, grout was injected in stages of30 centimeters in the vertical direction for the full2.8 meter depth which was required (betweenelevation -5.5 and elevation -8.3). By following aprimary and secondary injection pattern, a wall ofchemically grouted soils was installed around eachgroup of piles. see figure 6. Voids, caused by thedifference in rigidity between slabs, soils, pile caps,and the generation of settlements of the underlyingmud of the San Francisco Bay, were suspected to bein this area. Before injecting the chemical grout, aprobing program was adopted to look for voidsimmediately below both the basement floor slab andthe pile caps. Encountered voids were filled up andthe soils were tightened.

Aftercompleting the chemical injections, a lensingprogram for layer 1 was initiated. Lense groutinjections were spaced on a 1.S x I.c-meter grid,covering the space between the pile caps. Theseinjections were extended vertically between thebottom of the basement floor slab and the top oflayer 1, elevation -2.8 to elevation -5.5. A verticalstaging of 30 centimeters was used. In each stage28 liters of l2-sack cement slurry was injected (onesack = 42.7 kilograms). The initiation pressure wasbetween 700 and 3000 kPa, then dropped to 200 to450 kPa. There were few instances where thesepressures were not achieved" until several tens ofliters of grout were emplaced, indicating thepresence of a void.

3.3 Results and Discussions

Laboratory prepared samples of site sand withchemical grout, indicated an unconfined compressivest.rength well above the required 190 kPa at 28 days,see Table 1. Grout takes and pressures for eachinjection were checked to assure the properinstallation of each stage. During the primarychemical injections, the feedback pressure at thepoint ranged between 70 and 200 kPa. The

secondary injections ranged between 70 and 500kPa, indicating grout presence within the nearbysoils.

Table 1. Chemical Grout Strength

Sample

UnconfinedCompressiveStrength, kPaAge, Days

1234

27282830

276262283310

For the lense grouting program, no testing of thegrouted soils can be practically made. Qualityassurance was achieved by monitoring the groutamounts and pressures at the point of injection.

4 CONCLUSIONS

The risk of potential liquefaction can be eliminatedor at least reduced in certain cases usingcost-effective methods without interrupting thefunctions of the facility. The increased awarenessof owners, engineers, and public agencies to the soilliquefaction potential beneath their projects usuallymakes them search and explore available methodsfor an acceptable solution.

Geotechnical grouting techniques have beenavailable for quite some time (chemical grouting forabout 150.years, compaction grouting for about 30years, and lense grouting for about IS years), buttheir adaptation and use for the abatement of soilliquefaction has been limited to the last 10 to ISyears. The use of compaction. grouting inhorizontally driven casings proved to be a workablesolution. At least as far as the author is aware, suchan approach has never been tried before to theextent used in Case 1 above.

In Case 2, the use of a combination of groutingmethods other than compaction grouting wasdictated by the presence and condition of the timberpiles and the soil condi tions. The goals of bothprojects were successfully achieved to the pointwhere the subject soils were made non-liquefiableunder design-earthquake events.

REFERENCES

Al-Alusi, I-LR. 1994. Soil improvement to mitigatesettlements under existing structures.Proceedings of Settlement '94. Vertical andHorizontal Deformations of Foundations andEmbankments, Geotechnical EngineeringDivision, American Society of Civil Engineers,June 16-18, College Station, Texas

Bowles, lE. 1982. Foundation Analysis andDesign. McGraw-Hill Publishers, Inc.

Harding Lawson Associates 1990 and 1992.Geotechnical Investigation Reports for TapiaWastewater Reclamation Plant Expansion.

Holeyman, A. & Wallays, M. 1984. DeepCompaction by Ramming (in French).Proceedings of International Conference In-SituSoil and Rock Reinforcement, Paris, 367-372.

Mitchell, J.K. & Wentz, F.J., Jf. 1991. Performanceof improved ground during the Lama PrietaEarthquake. Report No. UCBJEERC-91/12,Earthquake Engineering Research Center,University of California at Berkeley.

Tokimatsu, K. & Seed, H.E. 1987. Evaluation ofsettlements in sands due to earthquake shaking.Journal of the Geotechnical EngineeringDivision, Proceedings of the ASCE, Vol. 113,No.8, August, 861-878.

Winterkorn & Fang 1991. Foundation EngineeringHandbook. Van Nostrand Reinhold.