8.6 Cofferdams, Underpinning & Adjacent Infrastructure Effects

Cofferdam Types and Function

A cofferdam is a temporary watertight (or substantially water-resistant) enclosure that permits excavation, foundation construction, or repair work to proceed in the dry, typically at a bridge pier, waterfront structure, or below-grade facility with a high groundwater table.

• Braced sheet-pile cofferdams: Single-wall interlocking steel sheet piles driven around the work area and internally braced with walers and struts (or externally anchored), analogous in structural behavior to a braced excavation but designed additionally for hydrostatic and wave/current loading on the exposed wall.
• Cellular cofferdams: Used for larger, deeper, or open-water applications (e.g., dry docks, large pier foundations) where internal bracing is impractical. Interlocking sheet piles are driven into circular cells (independent, self-stable circular cells connected by smaller connecting arcs) or diaphragm cells (straight diaphragm walls connecting circular arcs, used for elongated footprints), then filled with compacted granular material. Stability derives from the cell fill's internal friction and unit weight resisting overturning and sliding — much like a gravity retaining structure — while sheet-pile interlock tension must be checked against manufacturer allowable interlock strength.
• Double-wall cofferdams: Two parallel rows of sheet piles tied together and filled with granular soil between them, used where a single cellular cofferdam's minimum practical diameter cannot fit the site.

Unwatering and Bottom Stability

Once enclosed, the cofferdam interior is pumped dry ("unwatered"). Two geotechnical failure modes govern bottom stability when a hydraulic head differential exists across the embedded sheet-pile toe:

Piping/heave beneath the toe: Seepage flows beneath and around the embedded sheet-pile cutoff from the high-head (river/bay) side toward the low-head (dry excavation) side. If the upward exit gradient at the excavation-side toe approaches the soil's critical gradient, a "sand boil" or general heave/piping failure can occur, rapidly flooding the enclosure. A simplified screening check uses:

$FS_{piping} = \frac{i_{cr}}{i_{exit}}, \qquad i_{cr} = \frac{G_s - 1}{1+e}$

with the exit gradient approximated for a single sheet-pile cutoff of embedment depth $D$ beneath the excavation base and differential head $\Delta h$ as $i_{exit} \approx \Delta h / (2D)$ (a simplified flow-net approximation; project-specific work uses a full flow-net or seepage software solution). A minimum $FS \ge 1.5$ is typically targeted; if the check fails, embedment is increased, relief wells are installed on the dry side to reduce excess head, or a positive cutoff into an underlying low-permeability stratum is pursued.

General basal heave (cohesive soils): Where the cofferdam bottom is in soft clay, unrelieved weight of the retained soil/water column outside the cofferdam can push the excavation bottom upward, following the same bearing-capacity-based heave mechanism used for deep braced excavations — checked against the retained soil's undrained shear strength and the excavation depth/width geometry.

Underpinning Methods

Underpinning transfers an existing foundation's load path to a new, deeper, or higher-capacity bearing element, typically to permit adjacent excavation, correct settlement/distress, or accommodate increased loads.

Settlement Risk to Adjacent Infrastructure

Nearby excavation, dewatering, cofferdam installation, or underpinning sequencing can induce settlement in adjacent structures through ground loss (sheet-pile installation/extraction, jet-grout spoil return), stress relief (excavation unloading), or dewatering-induced consolidation. Risk is managed by:

• Defining a zone of influence around the work (often approximated as extending outward from the excavation base at roughly $1\text{H}{:}1\text{V}$, refined per project soil conditions) and identifying all structures/utilities within it.
• Establishing angular distortion and differential settlement limits for each affected structure — commonly $1/500$ for masonry-bearing-wall structures and looser limits (e.g., $1/300$) for steel/concrete frames — as monitoring "alert" and "limit" thresholds, not just a post-construction check.
• Continuous instrumentation: precise optical/laser survey of settlement points, tiltmeters on facades, crack monitors across existing distress, and (for dewatering-driven risk) piezometers to confirm drawdown stays within predicted limits.
• Sequencing underpinning or jet-grout work in short, alternating, non-adjacent segments to avoid simultaneously de-supporting long lengths of an existing footing.

Worked Numerical Example: Cofferdam Piping Factor of Safety

Problem Statement A braced sheet-pile cofferdam is installed for a bridge pier footing in a river with a design differential head of $\Delta h = 6\text{ m}$ between the river water surface and the dry excavation floor. The sheet piles are embedded $D = 4\text{ m}$ below the excavation base into a medium-dense sand with $G_s = 2.65$ and void ratio $e = 0.60$. Evaluate the factor of safety against piping at the toe.

Solution

Step 1 — Critical hydraulic gradient: $i_{cr} = \frac{G_s - 1}{1+e} = \frac{2.65-1}{1+0.60} = \frac{1.65}{1.60} = 1.03$

Step 2 — Approximate exit gradient at the toe: $i_{exit} \approx \frac{\Delta h}{2D} = \frac{6}{2 \times 4} = 0.75$

Step 3 — Factor of safety: $FS_{piping} = \frac{i_{cr}}{i_{exit}} = \frac{1.03}{0.75} = 1.37$

Step 4 — Evaluate: $FS = 1.37$ is below the typical target of $1.5$, indicating an unacceptable piping risk at the design embedment. Increasing embedment to $D = 5\text{ m}$ reduces $i_{exit}$ to $6/(2\times5) = 0.60$, raising $FS$ to $1.03/0.60 = 1.72 \ge 1.5$ — an acceptable design revision — or, alternatively, relief wells could be installed on the excavation side to lower the effective differential head.