Liquefaction assessment

How bridges get built · Site investigation and testing · Liquefaction assessment

Also called: cyclic resistance, seismic ground hazard

Liquefaction assessment — concept diagram — Concept diagram (TheBridgeEng)

How it is done

Screen for susceptibility: young alluvium, hydraulic fill and reclaimed land, groundwater depth, and grading and plasticity of the fines.
Set the design motion (peak acceleration and magnitude for the bridge's performance level); run a site response analysis where soft deep soils change the motion.
Characterise each layer with CPTu soundings, SPTs with fines content and Atterberg limits on samples, and shear-wave velocity where needed.
Compute the cyclic stress ratio with depth and the cyclic resistance from corrected cone resistance or N-values, with one consistent procedure including magnitude and overburden corrections.
Assess the consequences: post-liquefaction settlement, residual strength, lateral spreading toward channels and slopes, downdrag on piles, and stability of abutments and approach fills.
Design piles for kinematic loads, loss of lateral support and lateral spread forces through liquefied layers, or treat the ground (vibro-compaction, stone columns, DCM, drains).
Verify any treatment with post-treatment CPTs against the target cyclic resistance.

Key numbers

Simplified procedureCSR = 0.65 (amax/g)(σv0/σ'v0) rd; factor of safety = CRR/CSR.
Fine-grained soils with PI of 7 or more generally behave as clay rather than sand (Boulanger and Idriss, 2006).
Assessment depth commonly 20 m below ground; deeper layers need overburden corrections and are rarely critical.
Post-liquefaction volumetric strainabout 1–5 % of layer thickness in loose clean sand.
Common target after mitigationfactor of safety of 1.1–1.3 in the treated zone.

Where it fits

Moderate and high seismic zones with young alluvium, reclaimed fill, estuaries and river banks.

Where it does not

Ignoring thin liquefiable seams under footings; they matter more than deep ones.

Choosing it

Use CPTu as the primary tool and SPT with fines content as a check in critical layers: the CPT is continuous and finds thin seams.
Treat the ground when abutments and approach fills would spread; design the piles through the layer when only the piers are affected and the layer is thin.
Prefer a site-specific response analysis to code spectra on soft, deep sites, where amplification can change the cyclic stress either way.
Prefer cyclic laboratory tests on high-quality samples for silts and intermediate soils, where CPT and SPT correlations are least reliable.

Plant, pace and money

PlantNone on site; it is analysis on the investigation data.

ProductivityDays of engineering.

CostSmall; the consequences it prevents are not.

Risks and controls

What goes wrong

Fines corrections are the largest uncertainty; use both SPT and CPT where the layer is critical.

Quality assurance

Independent check of the design motion and of the corrections applied.

Origins

Seed and Idriss published the simplified procedure in 1971 after the 1964 Niigata and Alaska earthquakes; CPT-based methods (Robertson and Wride, 1998), the NCEER consensus (Youd et al., 2001) and Idriss and Boulanger (2008, 2014) refined it into current practice.

Examples

Niigata 1964 (Showa Bridge) is the classic liquefaction failure of a bridge. At Kobe in 1995 liquefaction mainly damaged structures on reclaimed land, while the best-known viaduct collapses were structural column failures. The game raises seismic zones on estuary and reclaimed sites.

Case studies

Showa Bridge, Niigata earthquakeJapan · 1964Lateral spreading of liquefied riverbed sand displaced the pile bents and five simply supported spans fell, a month after the bridge opened: the classic liquefaction failure.

Hyogo-ken Nanbu (Kobe) earthquakeJapan · 1995Reclaimed islands liquefied widely and quay walls moved several metres seaward; on the waterfront, ground movement displaced bridge piers and an approach span of the Nishinomiya-ko Bridge fell.

Canterbury earthquakesNew Zealand · 2010–2011Lateral spreading along the Avon River pushed abutments of short-span bridges toward the channel, rotating them while the decks acted as struts; it is now a reference case for abutment design in spreading ground.