Diaphragm Wall Design

Nicholson Construction Company 12 McClane Street Cuddy, PA 15031 Telephone: 412-221-4500 Facsimile: 412-221-3127

Diaphragm Walls by

Thomas D. Richards, Jr. P.E. Nicholson Construction Company, Cuddy, Pennsylvania

Presented at: Central PA Geotechnical Conference Hershey, Pennsylvania March 23-25, 2006

05-01-145

Diaphragm Walls Thomas D Richards, Jr P.E. Nicholson Construction Company

Central PA Geotechnical Conference - March 23-25, 2005

INTRODUCTION

The purpose of this paper is to describe the application, construction process, and design methods for diaphragm walls, since this topic has not been addressed much if at all at previous Hershey conferences. Diaphragm walls are a method of creating a cast in-situ reinforced concrete retaining wall using the slurry supported trench method. As such, they are often known as slurry walls. However, the term “diaphragm walls” Concrete diaphragm slurry walls were first introduced in the United States in the 1960s, and have found a niche in urban environments such as Boston, New York City, and Washington, DC.

APPLICATIONS

Diaphragm walls are most commonly used : •

in areas with dense and historic urban infrastructure,

•

where a very rigid earth retention system is required,

•

where noise and vibration must be limited,

•

where the geology and groundwater preclude the use of conventional earth retention systems

•

and/or where dewatering is not practical

Compared to other wall types, diaphragm walls are considered to be very stiff with respect to ground movement control (Clough and O’Rourke, 1990). Diaphragm walls are often attractive in granular soils with a high groundwater level, especially when a low permeability layer underlies the granular soils. The diaphragm walls are typically terminated in the underlying low-permeability layer which can consist of soil or rock. Keying into this low permeability layer reduces groundwater seepage below the wall. (Pearlman, 2004) Projects that have used these walls include: •

below grade parking/ deep basements

•

cut and cover subway tunnels

•

highways as cut and cover tunnel walls and for underpasses

•

shafts for deep sewers

•

dam appurtenances

•

landslides

For highway projects, diaphragm walls were employed extensively on the Central Artery Tunnel and also have been used in Denver, CO and Baltimore, MD.

Diaphragm Walls st 21 Central PA Geotechnical Conference - March 23-25, 2005

Page 2 of 17

BENEFITS

Diaphragm walls can: •

be formed to depths of several hundred feet, through virtually all soil types and through rock, and with great control over geometry and continuity

•

facilitate excavations below groundwater while eliminating dewatering

•

provide fairly watertight walls

•

•

provide structural stiffness which reduces ground movements and adjacent settlements during excavation be load bearing transferring loads to the underlying layer be reinforced to allow incorporation of many structural configurations,

•

accommodate connections to structures

•

be easily adapted to both anchors and internal structural bracing systems

•

be constructed in relatively low headroom (say 15 feet) and in areas of restricted access

•

be installed before excavation commences

•

provide economic solutions in cases where temporary and permanent support can be integrated or redesigned into one retaining structure

Diaphragm walls combine into a single foundation unit the functions of temporary shoring, permanent basement walls, hydraulic (groundwater) cutoff, and vertical support elements. Because of this combination, they have proven to be an economical alternative in many circumstances (Pearlman, 2004).


Page 3 of 17

CONSTRUCTION PROCESS Overview

The trench excavation is performed using slurry for support. The slurry is typically bentonite and water or polymer and water. Diaphragm walls are constructed in the following steps: •

pretrenching to remove obstructions

•

guidewall construction

•

panel (vertical segments) excavation

•

endstop placement

•

panel desanding

•

reinforcing cage placement

•

tremie concrete

•

end stop removal (if temporary)

Excavation

Cage Placement

Tremie Concrete

Figure 1 Source: George Tamaro, Mueser Rutledge


Page 4 of 17

Site Logisitics and Slurry Plant Setup

It is important to note that diaphragm wall installation requires sufficient work area to set up the slurry plant and to assemble the reinforcing cages prior to placement in the wall. This work may be difficult on congested sites. To reduce site area requirements, offsite cage fabrication is possible.

Cage Fab * ~120’ x panel depth

Slurry Plant with 6 tanks ** ~60’ x 120’

Cage Fabrication Area – Another Job

* The cage fabrication area is dependant on the number of rigs and production schedule. ** The plant are is dependant on number of tanks. The slurry plant includes a slurry mixer, storage tanks, and desanding units. Sufficient storage tanks must be used for bentonite slurry hydration, several panels of bentonite, recycled bentonite.


Page 5 of 17

Storage Tanks

Desanders and Desilter

Pretrenching

Pretrenching is often performed to remove shallow obstructions and provide stable support for the guidewalls (next step). This pretrenching may be performed as open excavation backfilled with flowfill or excavated under self hardening slurry.

Guidewall construction

Guidewalls provide a template for wall excavation and panel layout, support the top of the trench, restrain the endstops, serve as a platform to hang the reinforcement, provide a reference elevation


Page 6 of 17

for inserts ( anchors, slabs, etc.), support the tremie pipes, hold down the cage during concreting, and provide reaction for jacking out some types of endstops. Guidewalls are reinforced concrete typically four to five feet deep and constructed similar to the figure and photo below. The top of the guidewalls should be at least four feet above the groundwater table to allow for construction in the dry and to allow for slurry level to be three feet above groundwater table.

Typical Guidewall Details


Page 7 of 17

Panel (vertical segments) Excavation

Special clamshells also known as grabs or buckets are rectangular shaped (see photos) and used to excavate vertical slots known as panels. These clamshells may be cable hug or Kelly mounted, and the digging mechanics may be cable or hydraulic operated.

Kelly Mount Hydraulic Grab

Cable Mounted Hydraulic Grab

Cable Mounted & Operated Grab

Panel Excavation along Building


Page 8 of 17

The excavation is performed in “panels” which are vertical slots. Trench stability is mostly provided by the fluid weight of the bentonite and the arching action of the soil around the trench. Calculations on trench stability often do not show that successfully excavated trenches should stay open which indicates conservatism and effects that have not been considered. The bentonite slurry is placed in the trench after a few buckets have been excavated and continuously added to maintain at least 3 feet above groundwater level and within 2 feet of the top of the guidewall. Panel lengths are typically 20 to 24 feet governed by the geometry of the project and the size of contractors special clamshells. The panel width is governed by the contractors clamshells. Various widths can be accommodated by reinforcing design including shear and bending reinforcement.

Endstop Placement

Endstops are used to control the concrete placement so that adjacent secondary panels are not excavating monolithic concrete. Endstops may be permanent or removed after concrete placement. Permanent endstops are typically wide flange shapes. Removable endstops can be pipe (Figure 1) or special keyway end stops (Photo below).

Permanent Endstops

Special V Groove Endstop


Page 9 of 17

Panel Desanding

The panel must be de-sanded to remove excess sand in the slurry and bottom of panel. The removal of sand from the slurry decreases the density of the slurry so that tremie concrete does not mix with the slurry or trap pockets of sand. Reinforcing Cage placement

Carefully fabricated three-dimensional reinforcing cage are then inserted into the panel excavation. The reinforcing cage may also support future structural or utility connections using “knockouts” that are pre-set in the wall. Concrete is then placed around the reinforcing cage using tremie methods to form each concrete panel.

Cage Placement Note blockouts for floor slabs and trumpets for anchors .

Tremie Concrete

Tremie pipes are placed in the panel to within a foot of the bottom. Typically two tremie pipes are used for full size panels and one tremie pipe is used for single bite panels. Concrete with 8 to 10 inch slump is then tremied into the panel. The concrete mix is special to provide 4000 to 6000 psi strength with high slump and contains fairly high cement content, often other pozzolans, plasictizers and often other chemicals.


Page 10 of 17

The concrete level is sounded after each load and records maintained on actual versus theoretical concrete take. Tremie pipe sections are removed as the concrete level rises but maintained 10 feet into the concrete. While the concrete is being placed, the bentonite slurry is pumped back to storage tanks for treatment and reuse.


Page 11 of 17

End Stop Removal (if temporary)

As the concrete is setting typically four hours after placement at a given depth, temporary endstops are removed by crane or jacks (see Photo of Special V Groove End Stop above). This often means late nights and overtime.

DESIGN (modified from Pearlman, Boscardin, and Walker 2004)

The design analyses for excavation support systems can range from relatively simple empirical analyses to more complex computer analyses, where typically all stages of the excavation sequence are evaluated. The design considerations should include not only the stresses and loads on the support system, but also the affect of construction movements on the response of adjacent structures. The level of effort for the evaluation often depends on the stage of the project, proximity of structures, contractor’s methods of construction, and known local practice. The discussion of design methodologies will consider both structure loading and system movements. Empirical Methods Stress Analysis

Traditionally, apparent pressure envelope methods have been used successfully to design flexible wall systems such as soldier pile and lagging and steel sheet-pile systems. The approach was developed based on data from flexible wall systems, and typically assumes that the wall acts as a simple beam spanning between the brace levels (Terzaghi et al., 1996). For the more rigid slurry wall system, the pattern of wall displacement that develops during the actual excavation and bracing sequence can have a major effect on the bending moments in the wall and the distribution of load to the bracing/anchors. Hence, use of apparent pressure envelopes for design of stiffer systems can be misleading. In general, apparent pressure envelope loadings are most appropriate as upper bounds for cases that match the bases of the empirical data, which include cases with relatively flexible walls and a stable subgrade. The pressure envelope design approach is for a temporary support system and does not necessarily provide the long-term loading corresponding to the permanent condition after the end of excavation. When the temporary support system, such as a slurry wall system, is incorporated into the permanent building foundation, a staged analysis that includes loading at each stage is required to evaluate the built-in stresses and strains that are locked into the final structure at the end of construction. Movement Analysis

The use of empirical data for the evaluations of movements is a useful tool in evaluating potential effects of a proposed excavation on adjacent buildings. Empirical data also allow the designer to validate the general magnitudes and patterns of the results of more sophisticated analyses. The empirical data can be used to estimate the zone of influence of the excavation as well as typical magnitudes of ground movements for various wall stiffness and subgrade stability conditions (e.g., Clough and O’Rourke, 1990).


Page 12 of 17

Staged Excavation Analysis

Staged excavation analyses use numerical approaches to model the actual sequence of excavation and brace installation by considering each stage of the excavation as it is constructed, and the excavation support is installed and then removed. The soil and water pressures applied to the wall are representative of the actual pressures (not apparent pressure envelopes) expected in the system at each stage, and calculated loads are representative of the actual loads (not upper bound loads). The models can incorporate interaction of the soil and the structure as the earth pressures vary with displacement. The overall reliability of the structural requirements and displacement performance estimates determined from a staged excavation analysis is directly related to and very sensitive to the quality of the input parameters, particularly soil stiffness and strength parameters. Three general methods have been used for staged construction analyses: 

Equivalent Beam Method



Beam on Elastic Foundation Method



Finite Element Method

The equivalent beam method is outdated and rarely used in current practice. Discussion will focus on the beam on elastic foundation and finite element methods. Both approaches can be used to predict stresses, loads, and system movements. Beam on Elastic Foundation Method (BEF)

The earth pressures are modeled with a series of independent spring supports similar to Winkler elastic foundation model. At the start of the model, the springs are compressed to create an initial load equal to represent a state of at-rest pressure. At each stage of excavation or support system, the spring loads change as soil, water, and support system loads are applied or removed and lateral wall displacement occurs. The soil springs load-displacement relationship (modulus of subgrade reaction) is determined by the input soil stiffness and governs the spring displacement until the limiting value of active or passive pressure is reached. The Winkler elastic foundation model approximates the wall-soil interaction with a onedimensional model instead of a two-dimensional model that includes the soil mass, and hence does not include the effects of arching within the soil mass. Typically, the required soil parameters include: unit weight; at-rest, active, and passive earth pressure coefficients; and values for the modulus of subgrade reaction for the various soils that may affect the system. The modulus of subgrade reaction is not a true soil property, but rather depends on both the soil conditions and the geometry of the excavation being modeled. To be representative, the modulus of subgrade reaction needs to be adjusted based on the effective influence zone, which varies with the size of the loaded area. Typically, the predicted wall displacements are much more sensitive to the values of subgrade modulus used in the analysis than the predicted brace loads and wall bending moments. Hence, conservative selection of the modulus of subgrade values should provide conservative estimates of ground movements, without significantly increasing the structural demand of the wall and bracing system.


Page 13 of 17

The BEF method does not directly estimate vertical ground movements behind the wall. Ground movements behind the wall are evaluated using the calculated wall displacement from the model. An empirical relationship between wall movement and ground movements must then be used. There are several computer programs that automate the analysis. Some use Young’s modulus as input for the soil stiffness. The program then automatically converts the Young’s modulus values for the various soils to adjusted values of subgrade reaction modulus using closed-form elastic solutions. The BEF analytical model can provide useful insights into the behavior of the wall and the wall-soil boundary, and the automated computer programs make it easy to perform multiple analyses for optimizing the design and evaluating sensitivity to input parameters. Finite Element Methods (FE)

Finite element models are typically two-dimensional models that include the soil mass surrounding the excavation. The stress-strain response of the soil is represented by a mathematical soil model that can vary from a simple linear-elastic model to a complex nonlinear elasto-plastic model. The stress-strain response can be defined in terms of effective stresses or total stresses. The required input parameters depend on the soil model used. Generally, it is desirable to use a soil model that can model failure (plastic yield) when the soil strength is exceeded. In some problems, the ability to model volumetric changes in the soil (consolidation or dilation) may be important. A linear elastic, fully plastic Mohr-Coulomb soil model is often used. In this soil model, the soil acts linearly elastic until it reaches failure, defined by the Mohr-Coulomb criterion, where upon it becomes perfectly plastic. In contrast to the BEF analysis, the FE analysis can provide direct information on the ground movements outside of and inside the excavation (see Figure 2 below). It can also be used to model the soil-structure interaction response of nearby structures to the excavation-induced ground movements.


Page 14 of 17

Figure 2: Displacement Vectors from FEM Analysis Another difference between the FE and BEF methods is that variations in the soil stiffness (modulus) can have a greater effect on predicted loadings and movements due to the inclusion of soil arching in the FE model. FE models can be used to perform parametric studies to understand the relative effects of changes of parameters such as soil stiffness and excavation support stiffness and sequence on forces, stresses and displacements. They can also be used to estimate the absolute magnitudes and patterns of excavation support systems and ground movements which is much more difficult. A primary reason for the difficulty is the selection of reasonable stiffness values for the various materials that make up the soil mass. In general, values of stiffness based on laboratory and field tests tend to underestimate to a large degree the ground stiffness. This in turn can result in an overestimate of the magnitude of displacements, by two times or more, and the extent of the influence zone around an excavation. This tendency can be tempered to a great degree by using representative, local, field case history data during the selection of material parameters and to calibrate the numerical model to previous case histories. In the past, performing FE analyses have been complex and time consuming to perform, but new, user-friendly programs (e.g., PLAXIS) are making their use more common.


Page 15 of 17

Comparison of BEF and FE Results

For the United States Capital Visitor Center (Pearlman et.al. 2004 and Bonita (2005)), the analyses for the structural design of the support system were performed using both BEF and FE models. The BEF program (WALLAP, 1997) was easier and quicker to run than FE programs, so it was used for the structural design of the wall system. By using the BEF model, the design team could evaluate more design profiles. Two FE models were run to verify that the BEF model loadings and stresses were conservative, and to provide ground deformation predictions to compare to contract requirements. Figure 3 presents the predicted deflected shapes of the slurry wall for the BEF and FE model analyses, as well as inclinometer data for the most heavily loaded design sections. The FE model included the tieback anchors modeled within the soil mass. The difference between the movements predicted by the BEF model and the larger movements predicted by the FE model is essentially the free field movement behind the anchor zones of the tiebacks. In other words, the BEF and FE model had good agreement in predicting the local movement of the wall. The actual wall movement is less than the values predicted by both models. This behavior is likely the result of the combination of conservative modulus values for the soils, and conservative estimates of building surcharges used in the models. Pearlman et.al. (2004) note that overall the wall movements for the entire site are less than predicted, even in sections where there are no building surcharges.

Figure 3: Modeled and Measured Wall Displacement Data


Page 16 of 17

SUMMARY

Permanent retaining walls with high groundwater tables can be economically constructed using concrete diaphragm walls.

An introduction to construction methods was presented. Given the various options of permanent versus temporary endstops, panel length and width, and the economics of these options; the final design of the diaphragm wall is often done as design/build working in close cooperation with the Owner, GC, and Engineers.

Design techniques that involve sophisticated soil structure interaction models combined with local data and experience give a high level of confidence for predicting wall performance on projects surrounded by other structures, where control of building movement and damage are paramount to a successful project delivery. These models need to be calibrated to empirical predictions, and other case histories of successful excavation support projects in similar ground conditions.

ACKNOWLEDGEMENTS

Most of the design section of this paper was prepared By Seth Pearlman, Mike Walker and Marco Boscardin. REFERENCES st

Bonita, G. (2005) "United States Capital Visitor Center ", Proceedings of 21 Central PA Geotechnical Conference - March 23-25, 2005 Clough, W.G. and O'Rourke, T.D., 1990. "Construction induced movements of in-situ walls." Design and Performance of Earth Retaining Structures, ASCE GSP No.25, 439 - 470. PLAXIS, 1998. Finite Element Code for Soil and Rock Analyses. Brokgreve and Vermeer, et al., (ed.), Balkema. Rotterdam, Brookfield, Version 7, A.A. Pearlman, S.L., Boscardin, M.D., Walker, M.P. 2004. “Deep Underground Basements for Major Urban Building Construction,” Presented at Geo-Support 2004, Jan. 28-31, 2004, Orlando, FL. Terzaghi, K., Peck, R. B., and Mesri, G., 1996. Soil Mechanics in Engineering Practice, Third Edition, John Wiley & Sons, New York, NY, 349-360. WALLAP, 1997. Anchored and cantilevered retaining wall analysis program, D.L. Borin, MA, Ph.D., CEng., MICE. Geosolve, Users Manual, Version 4.


Page 17 of 17

Diaphragm Wall Design

Recommend Documents