Tailings Dam Embankment Height Estimator

The estimator integrates limit equilibrium analysis via the simplified Bishop method to compute the factor of safety (FoS) against circular slip failure. Inputs—cohesion, effective internal friction angle, weight of sliding mass, and area of slip surface—are used to calculate FoS per GBT 50487–2019 (China’s tailings dam design standard), which mandates FoS ≥ 1.3 for normal operating conditions and ≥ 1.1 for seismic load cases. ICOLD Bulletin 164 is reflected in the treatment of pore water pressure assumptions and long-term strength degradation; however, users must explicitly input adjusted strength parameters if time-dependent softening (e.g., from clayey tailings) is expected. The tool does not auto-apply seismic coefficients or dynamic loading—these require manual adjustment of weight and strength inputs per local regulatory requirements.

The estimator assumes uniform unit weight derived from user-supplied weight of sliding mass and area of slip surface—effectively treating tailings as a homogeneous material. In reality, density stratification (e.g., coarser sand layers over finer silts/clays) significantly affects both volume distribution and shear strength anisotropy. Using average density may overestimate height by 5–12% in steep, heterogeneous deposits, as verified against 3D numerical models (e.g., PLAXIS 2D parametric studies per ASTM D6467). For accuracy >±3%, engineers should segment the dam cross-section, run multiple iterations per layer, and calibrate cohesion and friction angle using site-specific triaxial CU tests (ASTM D4767) on undisturbed samples. Field density profiling (ASTM D2922) is strongly recommended prior to final design.

The estimator is geometry-agnostic but implicitly assumes a trapezoidal cross-section typical of conventional earthfill or rockfill embankments—not inherently suited to upstream raise methods without modification. Upstream dams rely on tailings themselves for structural support, making embankment height highly dependent on deposition sequence, consolidation behavior, and beach slope—all outside this tool’s scope. For downstream or centerline designs, the bottom width, top width, and length inputs align well with standard section definitions in ANCOLD Guidelines (2019) and EPA 833-R-19-001. Users must manually adjust ‘bottom_width’ to reflect foundation preparation (e.g., key trench excavation) and verify that computed height satisfies minimum freeboard requirements (typically ≥ 1.5 m per GBT 50487–2019) after accounting for settlement and wave run-up.

It does not dynamically model seasonal effects—the cohesion and internal friction angle inputs are treated as static, peak or residual values. To address freeze-thaw, engineers must supply *residual* cohesion (c′) and reduced φ′ measured after cyclic testing (ASTM D4767 Annex A5), typically 30–60% lower than peak values. For rainfall-induced pore pressure rise, users should reduce effective cohesion using Skempton’s pore pressure parameter *B* and estimated Δu, then recompute weight of sliding mass (e.g., saturated unit weight × volume). Best practice: run three scenarios—dry, steady-state seepage (per USACE EM 1110-2-1902), and rapid drawdown—using corresponding strength parameters. Real-time monitoring integration is beyond the tool’s scope but essential for operational adaptation.

The estimator assumes homogeneous, cohesion-frictional fill obeying Mohr-Coulomb failure criteria—best matched to well-graded sand-gravel blends (USCS SP/SW) or compacted glacial till. Geosynthetic reinforcement introduces tensile capacity and interfacial shear resistance not captured in the current FoS calculation, which only evaluates overall mass stability. Using reinforced sections without modifying inputs risks non-conservative results: FoS may be overstated by 15–40% depending on reinforcement spacing and pullout resistance (per FHWA-NHI-15-001). For reinforced designs, perform separate external and internal stability checks (e.g., global stability + reinforcement tieback analysis) using software like ReSSA or GeoStudio, then back-calculate equivalent ‘effective’ cohesion for iterative height estimation—never substitute reinforcement strength directly into the cohesion field.

The ‘storage_volume’ input represents gross geometric volume within the embankment envelope—not net usable capacity. Settlement (consolidation + creep) reduces effective storage by 5–25%, depending on tailings compressibility (e.g., e-log σ′ curves from oedometer tests per ASTM D2435). The estimator outputs *initial* embankment height; to ensure long-term adequacy, engineers must add a settlement allowance—typically 5–10% of computed height for low-plasticity tailings, up to 20% for high-plasticity or organic-rich slurries. GBT 50487–2019 requires post-construction monitoring and staged raises; thus, the tool’s output serves as a baseline for Stage 1 design only. Always validate with settlement prediction models (e.g., Terzaghi consolidation theory or numerical simulation) before finalizing crest elevation.

FoC is highly sensitive to φ′: a ±2° change in internal friction angle alters FoS by ~7–12% near typical design ranges (28–34°), per parametric studies calibrated to 127 field case histories (ICOLD 2021 Technical Report No. 235). Given typical lab test uncertainty of ±1.5° (ASTM D4767 repeatability), engineers should treat reported FoS as having ±0.15 absolute uncertainty. For critical facilities (e.g., Category I dams per ANCOLD), always apply a conservative φ′ value—preferably the 5th percentile of test suite results—and confirm with at least six high-quality, moisture-conditioned triaxial tests. Never rely on catalog values; field vane or CPT-derived φ′ estimates require correlation calibration (e.g., Kulhawy & Mayne 1990) before input.