Simulation of crest crack in high earth-core rockfill dam using XFEM 2-D modeling

Abstract Crest cracking represents a pervasive structural hazard during the construction and operational phases of high earth-core rockfill dams (ECRDs); consequently, the development of history-matched numerical tools is rendered an engineering necessity to prevent catastrophic risks. To address the limitations inherent in traditional, mesh-dependent techniques regarding the capture of discrete macro-fractures, an advanced numerical framework is introduced, wherein a hybrid Neural Network Genetic Algorithm (NNGA) is dynamically coupled with an Extended Finite Element Method (XFEM) solver. The fundamental novelty of this research lies in addressing a critical gap in contemporary literature, namely the historical decoupling of fracture mechanics from the concurrent effects of in-situ wetting deformation, long-term creep, and clay core consolidation. By executing a history-matched inversion of multi-year geodetic monitoring data from a 186-m-high ECRD, three primary contributions are delivered: (i) a site-calibrated, time-dependent material parameter baseline is established, thereby bridging the scale gap between laboratory testing and prototype realities; (ii) an explicit, physics-based quantification of crack initiation, propagation depth, and the long-term structural stabilisation timeline driven by the uncoordinated core–shell 'dragging effect' is provided; and (iii) a highly transferable predictive protocol for infrastructure assessment is introduced. Exceptional mathematical alignment with long-term field data is demonstrated by the quantitative validation results (overall R 2 = 0.994, RMSE = 0.037 m), whereby a maximum structural settlement of 3.27 m and the rapid propagation of a 1.75 m deep longitudinal crest crack—which structurally stabilises over a ten-year operational timeline—are accurately predicted. Ultimately, this integrated workflow is intended to serve as an invaluable asset for dam design engineers optimizing core–shell zoning configurations, field asset operators executing structural health monitoring, and safety regulators establishing scientifically backed risk-mitigation timelines for ultra-high embankment structures globally.