Bridge topics · Substructure

Integral vs Semi-Integral Abutment: Differences and Limits

Two ways to build a bridge without expansion joints at its ends. The integral abutment moves with the deck on flexible piles; the semi-integral abutment moves only a shallow end screen and keeps the deck on bearings.

Why get rid of the joints

Expansion joints and bearings at the ends of a bridge are where most durability trouble starts. Joints leak, de-icing salt and debris run through them onto the bearing shelf and the abutment face, steel corrodes, concrete spalls, and the joints themselves wear out long before the structure does. A British survey in 1989 found that most expansion joints inspected on existing bridges had failed. The response, first in North America and the United Kingdom and now worldwide, was the jointless bridge: make the deck continuous over its piers and connect it to its abutments so that thermal movement is absorbed by the substructure and the soil behind it instead of by a joint.

There are two main ways to do this at the abutment: integral and semi-integral.

Integral abutment

In an integral abutment the deck, the girder ends and the abutment wall are cast together into one monolithic piece. There are no bearings and no expansion joint at the abutment. The whole abutment moves with the deck as it expands and contracts with temperature, so it must be founded on something that can follow that movement without large restraint forces. In North American practice that is usually a single row of flexible piles, most often steel H-piles, embedded about half a metre into the abutment cap. The piles bend as the abutment moves back and forth; the soil behind the wall is pushed and released every season.

Typical features:

• girder ends encased in the end diaphragm and abutment wall, giving a moment connection;
• one row of piles, chosen and oriented for flexibility in the direction of movement;
• a short, shallow abutment so that the moving mass and the passive soil resistance stay small;
• a reinforced concrete approach slab tied to the deck, so that the joint moves to the far end of the approach slab, usually onto a sleeper slab (FHWA's design example uses a 25 ft, about 7.6 m, approach slab).

Semi-integral abutment

A semi-integral abutment keeps the benefit at the top and gives up the moment connection at the bottom. The deck ends in a deep end diaphragm, or end screen wall, cast integral with the girders, which retains the backfill and moves with the deck. Underneath, the deck still sits on bearings on a conventional, fixed abutment seat that does not move. The joint between the moving end screen and the fixed abutment is below the deck and protected from surface water, and there is still no expansion joint in the running surface at the end of the bridge.

This arrangement is used when the substructure cannot follow the deck: stiff foundations on rock, tall abutments, spread footings, or existing abutments being re-used under a new deck. It also suits longer bridges, because the bearings, not the piles, take the movement. The price is that bearings come back, with their inspection and eventual replacement, and the detail at the end screen needs care so that the backfill does not leak into the gap and water does not reach the bearing shelf.

Integral vs semi-integral at a glance

How long can a jointless bridge be?

The limit is not structural strength but how much movement the abutment, the soil and the approach can tolerate. Guidance differs by country and by owner:

• FHWA design example. Assuming about 100 mm (4 in) of total length change from extreme cold to extreme heat, that is 50 mm at each end, FHWA's LRFD design example arrives at roughly 180 m (600 ft) for concrete bridges and 120 m (400 ft) for steel bridges in moderate climates, and less in cold ones. Most US states publish their own length and skew limits, often more restrictive.
• United Kingdom. National Highways' CD 350 requires integral construction to be considered for bridges up to 60 m long with skew up to 30°. Movement at the top of the abutment is limited to about 20 mm, and the 40 mm range of the asphaltic plug joints used at the end of the approach puts a practical ceiling of around 80 m on the total length.
• Skew. There is no universal limit. Skew makes the abutment rotate in plan as it moves and raises soil pressures at the acute corner, so many owners cap integral bridges at 20° to 30°, and semi-integral bridges are rarely designed beyond about 20°.

Steel decks move more than concrete decks per degree of temperature change, so their limits are shorter. Concrete decks also shorten with creep and shrinkage, which helps in summer and hurts in winter.

The soil behind the wall

Every expansion cycle pushes the abutment into the backfill and every contraction lets it fall back. Over many cycles the soil densifies and the lateral pressure ratchets up well beyond the active pressure used for a conventional wall. UK practice designs for an enhanced earth pressure coefficient, K*, which grows with the movement of the wall and is held at its maximum value for the design expansion at the ultimate limit state. The practical defences are a short abutment, compressible or well-graded granular backfill, a drained back face and, where the movement is large, a semi-integral solution that moves only the shallow end screen.

Approach slabs

The approach slab carries traffic over the zone of backfill that moves and settles. Tied to the deck, it moves with the bridge and pushes the joint away from the structure, several metres back, where a sleeper slab and a joint or flexible pavement take up the movement. Approach slabs are usually cast on a slip layer such as polyethylene sheet. Some owners avoid them and accept a joint at the abutment end screen instead; others regard them as essential. Either way, the movement does not disappear: it has to be accommodated somewhere along the road.

When to choose which

• Integral: new bridges up to the length limit of the owner, on soils where a row of flexible piles is practical, with modest skew and abutment height.
• Semi-integral: rock or stiff foundations, tall abutments, longer decks, larger skews within the owner's limit, or deck replacement on existing abutments where the old seat can carry new bearings.
• Conventional seat abutment with joint: long bridges beyond both limits, or where very large movements, seismic displacements or settlement rule out the soil working with the structure.

Related on TheBridgeEng

Questions engineers ask

What is a semi-integral abutment?

An abutment where the deck ends in an end diaphragm or screen wall cast with the girders, which retains the backfill and moves with the deck, while the deck itself sits on bearings on a fixed abutment seat. There is no expansion joint in the road at the bridge end.

What is the difference between an integral and a semi-integral abutment?

In an integral abutment the whole abutment is monolithic with the deck and moves on flexible piles, with no bearings. In a semi-integral abutment only the end screen moves with the deck, and bearings on a fixed substructure take the movement.

How long can an integral bridge be?

It depends on the owner and the climate. FHWA's design example gives about 180 m for concrete and 120 m for steel in moderate climates; UK guidance requires integral construction to be considered up to 60 m and 30° skew. Many owners set their own limits.

Do integral bridges need approach slabs?

Most US owners use a reinforced approach slab tied to the deck, moving the joint to a sleeper slab at its far end. Some UK clients avoid approach slabs and detail the end of the bridge differently, but the movement must be accommodated somewhere.

Why do integral abutments use H-piles?

A single row of slender steel H-piles bends easily as the abutment moves, so the restraint forces on the deck and piles stay small. Pipe piles and concrete piles are also used where they are flexible enough.

Sources

• FHWA LRFD Steel Bridge Design Example: Integral Abutment Design
• Integral bridges (SteelConstruction.info)
• Integral bridge (Wikipedia)

Updated 2026-10-25 · written by a practising bridge engineer