An official website of the United States government
Earthquake-related failure modes for embankment dams are commonly evaluated through numerical simulations using finite element or finite difference approaches. This is especially true for liquefaction triggering or cyclic softening of fine-grained materials where advanced constitutive models are used to capture the dynamic response of the dam and the nonlinear behavior of the soil. Both liquefaction and cyclic softening can lead to significant strength loss, which can lead to large deformations within the dam, but these numerical tools often cannot capture these large deformations due to excessive mesh distortion and subsequent numerical errors. This leads to significant uncertainties in estimating potential crest settlement, which is often a critical value for risk assessments of dams. Hybrid numerical methods like the material point method (MPM) offer a promising alternative to model large deformations, but their application to dams is still limited and relatively little validation has been done on using MPM for post-earthquake stability analyses. This study focuses on applying MPM simulations to evaluate the post-earthquake stability of a hypothetical embankment dam and to examine potential deformations of a flowslide that occurred in Palu, Indonesia in 2018. The MPM program Anura3D is used for the analyses with modifications to allow for assigning residual strengths. The results from the Palu flowslide are compared with observations from the field to show that the MPM analyses are able to capture the extent of the slide, but underpredict the measured displacements in the central portion of the flowslide. The analyses for the embankment dam are compared with post-earthquake stability results from finite difference analyses using FLAC. The MPM analyses are able to capture the full deformation of the flowslide, while the FLAC analyses are halted due to excessive mesh deformation. These results demonstrate the potential of MPM to be used as a complement to existing numerical tools for evaluating the seismic response of dams, but additional work is needed to validate this approach using case histories with both large and small deformations.
more »
« less
This content will become publicly available on September 25, 2026
SEISMIC-INDUCED CRACKING OF EARTH DAMS: UPDATED DATABASE AND RECOMMENDATIONS
Earth dams, when subjected to seismic loads, may exhibit longitudinal and lateral deformations, settlement, and the formation of longitudinal and transverse cracks. Cracking poses a severe threat to these structures, as it may lead to piping failure due to increased seepage and internal erosion through the cracks. Ensuring the safety of earth dams relies on an adequate assessment of their seismically-induced deformations. Current empirical methods for estimating the size and depth of longitudinal and transverse cracking produced during an earthquake are grounded in case studies from the 1960s to the 1990s. This study expands and modernizes the existing database, with information on the performance of 385 dams during 21 different seismic events, from 2000 through 2023. Data collection involved an exhaustive search from existing databases, published reports of seismic damage on embankments and earth dams, and from publications from technical journals and conferences. Additionally, the correlations by Pells and Fell (2002), which relate the damage class with the seismic intensity of the earthquake (characterized by both magnitude and Peak Ground Acceleration), have been updated. The data gathered, together with new correlations may be used by designers to enhance the seismic resilience of embankments and earth dams, as well as by researchers to further our knowledge on the seismic response of dams, to develop new numerical models, or calibrate or verify existing ones.
more »
« less
- Award ID(s):
- 2226154
- PAR ID:
- 10659451
- Publisher / Repository:
- Association of Dam safety Officials https://damsafety.org
- Date Published:
- Format(s):
- Medium: X
- Location:
- Cleveland, OH
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
The breaching of the Edenville dam and overtopping of the Sanford dam on May 19, 2020 have been the focus of significant geotechnical investigation, while considerably less scientific focus has been directed towards documenting the resulting damage to highway embankments, bridges, and homes downstream of the dams. The Geotechnical Extreme Events Reconnaissance (GEER 2020) Association deployed a small team to conduct field reconnaissance after the event with the goal of making preliminary observations of impacts to geostructures downstream of the Edenville and Sanford dams. Damage was concentrated in the city of Sanford while in the city of Midland, downstream of Sanford, flood waters did not reach heights that caused significant damage to bridges or property. The observations made by the GEER team suggest that bridge abutments, buried utilities in earthen structures, and timber-framed buildings downstream of dams and other high risk flood zones warrant additional considerations during design to increase resilience during extreme events.more » « less
-
Abstract. In this paper, two different convolutional neural networks (CNNs) are applied on images for automated structural damage detection (SDD) in earthquake damaged structures and cracking localization (e.g., detection of cracks, their widths and distributions) at various scales, such as pixel level, object level, and structural level. The proposed method has two main steps: 1) diagnosis, and 2) localization of cracking or other damage. At first a residual CNN with transfer learning is employed to classify the damage in the structures and structural components. This step performs damage detection using two public datasets. The second step uses another CNN with U-Net structure to locate the cracking on low resolution images. The implementations using public and self-collected datasets show promising performance for a problem that had remained a challenge in the structure engineering field for a long time and indicate that the proposed approach can perform detection and localization of structural damage with an acceptable accuracy.more » « less
-
Abstract Repeated earthquake cycles produce topography, fault damage zones, and other geologic structures along faults. These geomorphic and structural features indicate the presence of co‐seismic permanent (inelastic) surface deformation, yet a long‐standing question in earthquake research is how much of the co‐seismic deformation field is elastic versus inelastic. These questions arise in part because it is unclear what measurable co‐seismic characteristics, such as off‐fault or distributed surface deformation and cracking, represent true unrecoverable deformation. One emerging descriptor of permanent co‐seismic deformation is surface strain magnitudes inferred from imaging geodesy observations. In this study, we present the surface strain field of the 2013 Mw7.7 Baluchistan strike‐slip earthquake in southern Pakistan. We invert co‐seismic displacement fields generated from pixel‐tracking of SPOT‐5 and WorldView optical imagery for co‐seismic surface horizontal strain tensors. We observe that co‐seismic strain field is dominated by negative dilatation strains, indicating that the co‐seismic fault zone contracted during the earthquake. We show that co‐seismic inelastic failure exhibits a relatively consistent width along the rupture that is localized to a zone 100–200 m wide on the hanging wall side. The width of co‐seismic permanent deformation does not correlate with variations in off‐fault deformation or surface geology. Based on comparisons to other recent earthquakes, we posit that the permanent surface strains reflect inelastic deformation of the faults inner damage zone, and that the width of this zone reflects fault maturity.more » « less
-
Abstract Post‐seismic debris flows are an important hazard following large earthquakes, propagating destruction downstream from hillslopes where coseismic landslides occur and extending damage for years after shaking stops. Data sets of post‐seismic debris flows are necessary to predict initiation and runout characteristics but are presently scarce. We used satellite imagery supplemented by field observations to compile an inventory of >1,000 debris flows associated with the 2015 Gorkha Earthquake in Nepal. We identified two distinct debris flow types: (1) Material from a coseismic landslide was remobilized in a steep channel during a later monsoon; and (2) a new post‐seismic hillslope failure occurred in saturated conditions and became fluidized and channelized. Runout distance was constrained by channel confluences and may be related to confluence geometry. Unstable landslide debris was largely flushed from steep channels during the first monsoon following the earthquake, and the rate of new hillslope failures tailed off over a few years.more » « less
