Research Paper
Reliability analysis of rapid drawdown of an earth dam using direct coupling

https://doi.org/10.1016/j.compgeo.2019.103336Get rights and content

Abstract

This paper combines seepage analysis with limit equilibrium analysis to investigate the safety of an earth dam over time and during drawdown. The numerical investigation is accomplished by directly coupling the deterministic software packages Seepage/W and Slope/W with the StRAnD reliability software. The first-order reliability method is employed in reliability analysis. Sensitivity analyses reveal that saturated hydraulic conductivity (ks), friction angle (ϕ′ ) and cohesion (c′) are the random parameters with the greatest contribution to failure probabilities. The cumulative effect of random saturated hydraulic conductivity (mainly) makes critical times and critical slip surfaces significantly different in the probabilistic and deterministic analyses.

Introduction

The stability of earth dams for long-term conditions depends on its geometry, material properties, and the forces to which the dam is subjected. The forces due to water are internal (pore-water pressures and seepage forces) and external (hydrostatic and hydrodynamic effects [1], [2]). During water discharge of an earth dam, the reduction of water level has two principal effects: modification of the internal Pore Water Pressure (PWP), and reduction of the external stabilizing hydrostatic pressure due to unloading. Therefore, safe operation of earth dams requires studies of seepage and stability for the upstream slope of the dam.

The level of water in a reservoir is affected by hydrological events and by user demand, in case of water supply dams. The partial or total emptying of the reservoir is part of a dam operation procedures. Partial or total emptying is required during maintenance and in case extreme events occur, such as earthquakes (stress changes due to earthquakes [3]), landslide into the reservoir or extreme rainfall. Dam equilibrium during the emptying procedure depends directly on the discharge velocity, with rapid discharges known to be critical. This leads to the problem known as Rapid DrawDown (RDD), with the discharge velocity (VRDD) being a critical parameter. The discharge velocity depends on characteristics of the hydraulic installations, such as the capacity of spillways.

The usual approach to RDD analysis is a transient seepage and limit equilibrium analyses using mean soil properties, to find deterministic factors of safety (FoS) [2], [4], [5]. However, the equilibrium depends on loading conditions and soil parameters, which can be modelled as random variables. In the last four decades, significant progress has been made into quantifying the variability of soil properties [6], [7], [8], and in performing reliability analysis of slopes and embankments in long term steady-state conditions [5], [9], [10], [11], [12], [13], [14], [15], [16], [17]. However, to the best of our knowledge and based on a comprehensive literature review, no studies were found addressing reliability of earth dams in transient conditions of rapid drawdown.

Different approaches are possible for probabilistic slope and dam stability analyses. Point estimate methods (PEM) are very efficient and popular in geotechnical engineering [18], [19], [20], [21] but were shown to be inaccurate for many problems by [22]. Accuracy can be increased by using higher-order moment methods [23], [24], but efficiency is lost. Approximation methods such as the first-order reliability method (FORM) are competitive with PEM in terms of computational cost, but this one is are more accurate and suitable for evaluating small (failure) probabilities at distribution tails. The FORM loses accuracy to Monte Carlo Simulation (MCS) only if the limit state functions are excessively nonlinear, in which case PEM also fails.

The equilibrium analysis of earth dams is based on locating the critical slip surface, i.e., the surface with the smallest safety factor. In general, only a limited number of slip surfaces is checked [25], [26], [27], [28]. The critical slip surface, in a probabilistic analysis, is generally believed to correspond to the minimum safety factor surface in the deterministic approach [29]. Likewise, one could assume that the critical time, in a probabilistic transient analysis, would be the same critical time of a deterministic analysis. These ideas are challenged herein the present paper.

This paper addresses the reliability analysis of an example earth dam during the rapid drawdown. The numerical study is conducted by the so-called Direct Coupling (DC) [30], [31], [32], [33], [34], [35], [36] between a deterministic software for slope stability (GeoStudio 2018 [37], [38]), and a structural reliability software (StRAnD [39]). The Direct Coupling method is not usual in geotechnical literature because, due to historical trends, point estimate and response surface surrogate methods became more popular [36]. However, direct coupling is the rule when it comes to (super-) structure reliability analysis using numerical models. Direct coupling allows the full features of deterministic numerical software to be explored in probabilistic or reliability analyses. In direct coupling, the probabilistic solver treats the numerical software as a black-box: it defines specific values (outcomes) of the random variables for which the numerical solution needs to be computed.

In this paper, the First-Order Reliability Method (FORM) is employed for reliability analysis, although many other algorithms are available in the StRAnD software [39]. FORM is known to be very efficient, as it computes failure probabilities with a limited number of calls to the deterministic solver (GeoStudio). The FORM method is based on an interactive search for the so-called design point, at which the limit state function is linearized. FORM yields good results when the problem is not excessively non-linear, as is the case for the transient dam equilibrium problem addressed herein. Searching for the design point requires gradients of the limit state function, with respect to random variables. These are evaluated by finite differences in the current implementation [40].

The remainder of this paper is organized as follows. The problem setting is presented in Section 2, which also briefly describes the performance function and reliability analysis techniques employed herein. The complete direct coupling technique for transient analysis is described in Section 3. Results of the transient numerical dam equilibrium analysis are presented and discussed in Section 4. Concluding remarks are presented in Section 5.

Section snippets

Review of current drawdown standards

There are different criteria for addressing the risk of Rapid DrawDown (RDD) in earth dams. In different codes, the discharge is referenced as: (i) a proportion of dam height; (ii) a proportion of dam volume; and (iii) a function of height per day [41], [42].

The RDD criteria as a function of dam height proportion (i) recommend drawdown analyses varying between 10% and 90% of the height, within 7 days and 4 months [43], [44]. The velocity of RDD as proportion of dam height is divided into hazard

Steps of the probabilistic modeling

Deterministic and probabilistic calculations were implemented by the coupling GeoStudio 2018 [37], [38] and StRAnD 1.07 [39] software. Details of the coupling approach are presented and discussed in Siacara et al. [40]. The algorithms implemented in Siacara et al. [40] were modified herein to evaluate earth dams in different long-term steady-state and transient conditions. The factors of safety (FoS) are evaluated using different limit equilibrium methods (e.g. Morgenstern and Price). The

Reliability Analyses and Results

Reliability analyses were performed for the same scenarios described in the mean value (deterministic) analysis, including four values of VRDD (VA = 2.0, VB = 1.5, VC = 1.0 and VD = 0.5 m/day). Reliability indexes (β) and probabilities of failure (Pf) where evaluated by the Direct Coupling GeoStudio and StRAnD software packages, as described in 2 Problem Setting, 3 Methodology. The limit state gradients were estimated by finite differences. The search for the DP was performed with the HLRF

Concluding Remarks

This paper addressed the transient reliability analysis of earth dams during rapid drawdown operations. Seepage and equilibrium analyses were performed numerically using the deterministic GeoStudio 2018 software. Reliability analyses were performed by the first-order reliability method in StRAnD software. A direct coupling technique was employed, in which the probabilistic software calls the deterministic software as a black box. An example earth dam was studied, and four different discharge

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

The authors thank the financial support by the Coordination for the Improvement of Higher Education Personnel (CAPES) for research funding (Grant number n° 88882.145758/2017-01) and the Brazilian National Council of Scientific and Technological Development (CNPq).

References (101)

  • H.M. Kroetz et al.

    Reliability of tunnel lining design using the Hyperstatic Reaction Method

    Tunn Undergr Sp Technol

    (2018)
  • R.Y. Liang et al.

    A reliability based approach for evaluating the slope stability of embankment dams

    Eng Geol

    (1999)
  • A.T. Beck et al.

    Strategies for finding the design point under bounded random variables

    Struct Saf

    (2016)
  • B.K. Low et al.

    Reliability analysis using object-oriented constrained optimization

    Struct Saf

    (2004)
  • D. Val et al.

    Optimization problem solution in reliability analysis of reinforced concrete structures

    Comput Struct

    (1996)
  • R. Rackwitz et al.

    Structural reliability under combined random load sequences

    Comput Struct

    (1978)
  • P.-L. Liu et al.

    Optimization algorithms for structural reliability

    Struct Saf

    (1991)
  • P.A. Lane et al.

    Assessment of stability of slopes under drawdown conditions

    J Geotech Geoenviron Eng

    (2000)
  • C. Viratjandr et al.

    Limit analysis of submerged slopes subjected to water drawdown

    Can Geotech J

    (2006)
  • K. Song et al.

    Effect of hydraulic properties of soil and fluctuation velocity of reservoir water on landslide stability

    Environ Earth Sci

    (2015)
  • K.-K. Phoon et al.

    Characterization of geotechnical variability

    Can Geotech J

    (1999)
  • K.-K. Phoon

    Reliability-based design in geotechnical engineering: computations and applications

    (2008)
  • K.-K. Phoon et al.

    Risk and reliability in geotechnical engineering

    (2015)
  • E.E. Alonso

    Risk analysis of slopes and its application to slopes in Canadian sensitive clays

    Géotechnique

    (1976)
  • K.S. Li et al.

    Probabilistic design of slopes

    Can Geotech J

    (1987)
  • J.T. Christian et al.

    Reliability applied to slope stability analysis

    J Geotech Eng

    (1994)
  • J.M. Duncan

    Factors of safety and reliability in geotechnical engineering

    J Geotech Geoenviron Eng

    (2000)
  • S. Lacasse et al.

    Risk assessment and dams - Recent developments and applications

    XVII ECSMGE, Geotech. Eng. Found. Futur.

    (2019)
  • N. Poddar et al.

    Deterministic and probabilistic stability analysis of earth slopes during rapid drawdown. 50th Indian Geotech

    (2016)
  • M.E. Harr

    Accounting for certainty

    Found Eng Face Uncertainty, Geo-Congress

    (2013)
  • E. Rosenblueth

    Point estimates for probability moments

    Proc Natl Acad Sci

    (1975)
  • Y.-G. Zhao et al.

    New point estimates for probability moments

    J Eng Mech

    (2000)
  • M.E. Harr

    Reliability based design in civil engineering

    (1987)
  • W.H. Tang et al.

    Probability-based short term design of soil slopes

    Can Geotech J

    (1976)
  • A.M. Hassan et al.

    Search algorithm for minimum reliability index of earth slopes

    J Geotech Geoenviron Eng

    (1999)
  • H. El-Ramly et al.

    Probabilistic slope stability analysis for practice

    Can Geotech J

    (2002)
  • Y. Wang et al.

    Practical reliability analysis of slope stability by advanced Monte Carlo simulations in a spreadsheet

    Can Geotech J

    (2011)
  • Y.M. Cheng et al.

    Simplified approach for locating the critical probabilistic slip surface in limit equilibrium analysis

    Nat Hazards Earth Syst Sci

    (2015)
  • B. Sudret et al.

    Stochastic finite element methods and reliability

    A state-of-the-art-report

    (2000)
  • M. Berveiller et al.

    Presentation of two methods for computing the response coefficients in stochastic finite element analysis

    9th ASCE Spec. Conf. Probabilistic Mech. Struct. Reliab., Albuquerque, USA

    (2004)
  • M. Berveiller et al.

    Stochastic finite element: a non intrusive approach by regression

    Eur J Comput Mech

    (2006)
  • Napa-garcía GF. Reliability Analysis of Rock Slope with Planar Failure Using the Direct Coupling Approach. ISRM Conf....
  • GeoStudio. Heat and Mass Transfer Modeling with GeoStudio. GEO-SLOPE. Calgary, AB, Canada T2P: GEO-SLOPE International...
  • GeoStudio. Stability Modeling with GeoStudio. GEO-SLOPE. Calgary, AB, Canada T2P 0T8: GEO-SLOPE International Ltd;...
  • Beck AT. Structural Risk Analysis and Design - StRAnD. Release 1. Sao Paulo, Brazil: Department of Structural...
  • AT Siacara et al.

    Reliability analysis of earth dams via direct coupling

    To Appear J Rock Mech Geotech Eng (Accepted Pap)

    (2020)
  • Environment Agency. Guide to drawdown capacity for reservoir safety and emergency planning. vol. 1. Environment Agency,...
  • Environment Agency. Guide to drawdown capacity for reservoir safety and emergency planning. vol. 2. Environment Agency,...
  • USACE. PART 220 — Design criteria for dam and lake projects. In: U.S. Corps of Engineers – Department of the Army,...
  • H. Babbit et al.

    Emergency drawdown capability

  • Cited by (26)

    • A review of time-dependent reliability analyses of slopes: Research progress, influencing factors, and future research directions

      2022, Transportation Geotechnics
      Citation Excerpt :

      This is the hottest research topic with the most research results in the field of time-varying slope reliability. With respect to the time-varying factor resulting from the fluctuation of the reservoir water level, multiple scholars have studied the time-varying reliability of the slope stability by considering the uncertainties in the geotechnical strength parameters (e.g., the internal friction angle and cohesion) [59,67,41,42], the seepage characteristics (e.g., hydraulic parameters and the water head) [10], or the coupling of these [43,55,49,26]. Similarly, with respect to the time-varying factors derived from precipitation, multiple scholars have studied the uncertainty of the slope stability factors [50], including the geotechnical parameters and seepage influencing factors (e.g., hydraulic conductivity and rainfall intensity and duration) [6,22,44] or the coupling of the two [1,51,66,62,20,16,42].

    • Semi-quantitative risk analysis of high concrete face rockfill dams under seismic action

      2022, Structures
      Citation Excerpt :

      According to the results of related studies [26], γc is taken as 1 %. The center point method [27] was used to calculate the seismic deformation reliability of the dam body in this study. This method is advantageous because the real probability distribution of the fundamental variables is not needed, and the statistical parameters of the structural function can be directly used to represent the reliability index.

    • Stability evaluation of earth-rock dam reinforcement with new permeable polymer based on reliability method

      2022, Construction and Building Materials
      Citation Excerpt :

      Pang et al.[15] presented a new and efficient methodology that couples a recently developed generalized probability density evolution method (GPDEM) with a spectral representation random function method to assess the seismic reliability. Siacara et al.[16] combines seepage analysis with limit equilibrium analysis to investigate the reliability of an earth dam over time and during draw down. Uncertainty in geotechnical engineering mainly comes from spatial variability of soil mass.

    View all citing articles on Scopus
    View full text