  

2014

Modelling Internal Erosion Within An Embankment Dam Prior To Breaching

Vazquez Borragan, Alejandro Master Thesis 10/4/2014

Modelling Internal Erosion Within An Embankment Dam Prior To Breaching M. A. Vazquez ([email protected] , 850323-3432) KTH Royal Institute of Technology, Stockholm, SWEDEN . October 2014

ABSTRACT There are still uncertainties in the safety of existing embankment dams. For instance, the majority of embankment dams in Sweden were built between 1950s and 1970s, designed and constructed to standards that might be unacceptable nowadays. Particularly, Vattenfall’s records stated that 40% owned embankments dams developed sinkholes (Nilsson, 1999). Moreover, internal erosion and its failure mechanisms of initiation and development are still not fully understood (Bowles et al., 2013). Also, internal threats are difficult to detect and interpret even using new instrumentation techniques. The aim of this Master Thesis is to identify failure mechanisms of embankment dams prior to breaching and hence, verify the reliability of a risk analysis after the breaching of the dam. The methodology consisted of monitoring an embankment dam prone to fail by internal erosion mechanisms. Finally the results were modelled using FEM to identify the risk of internal erosion prior to breaching. Key words: Internal erosion, numerical modelling, monitoring, embankment dam, risk analysis, breaching, piping, glacial till, rock-fill, Global Backward Erosion, stress analysis, seepage, instrumentation, piezometers, laser scanner, deformations, filter, hyperbolic model.

1

“Sometimes… Barriers are created with a retaining function. However, by nature some retained parts react, finding the mechanisms to get its freedom back. Meaning that, a wall can be defeated in its action. …a failure can also be a Victory”. -Alejandro Vazquez

“A veces… Las barreras se crean con fines de retención. Sin embargo, por naturaleza las partes contenidas reaccionan, Hallando los mecanismos que las liberan. Significando que, se puede vencer al muro en su acción. …un fallo también puede ser una Victoria”. -Alejandro Vázquez.

2

ACKOWLEDGEMENT First of all, I would like thank Johan Lagerlund, who has acted on behalf of Vattenfall Vattekraft and Vattenfall R&D to put his confidence on my skills and knowledge. I really appreciate the interest, time and effort he invested, which it has greatly helped to the success of this Master Thesis, and also to my professional development. I am also very pleased with all the members of the staff working at the Vattenfall R&D facilities in Älvkarleby, especially with Martin Rosenqvist, James Yang and Pelle Enegren(Master Thesis student), who showed a great interest and support during the performance of the experiments. Second of all, I would like to thank Stefan Larsson (KTH), especially, for all the advice provided to perform the academic writing and orientation to accomplish this Master Thesis, but also for being the link between Vattenfall and the university. Finally, and not for that the least importance, I wish to thank first my brother for his unconditional support; and to my friends for their support and motivation. P.D: This work is dedicated to those who unfortunately I am not able to say thank you physically: {Maria Pilar Vazquez Borragan (Susana, mother) (1955-2008)}, {Eugenio Vazquez Añon (Abuelo, grandfather)(1912-2009}.

Alejandro Vazquez Borragan Stockholm, Sweden. 2014

Sponsors:

3

Contents

ABSTRACT................................................................................................................................................ 1 ACKOWLEDGEMENT ............................................................................................................................... 3 INTRODUCTION ....................................................................................................................................... 7 State-of-art and limitations................................................................................................................. 7 Summary ............................................................................................................................................. 9 METHODOLOGY ...................................................................................................................................... 9 Dam design and materials properties ............................................................................................... 10 Building the dam ............................................................................................................................... 11 Instrumentation ................................................................................................................................ 14 Soil testing......................................................................................................................................... 14 Monitoring ........................................................................................................................................ 14 Failure Mode Analysis ....................................................................................................................... 16 Dam Failure ....................................................................................................................................... 20 RESULTS ................................................................................................................................................ 22 Dam design ....................................................................................................................................... 22 Slope stability analysis .................................................................................................................. 22 Monitoring and modelling internal erosion ..................................................................................... 24 Seepage monitoring ...................................................................................................................... 24 Built model (SVFlux) ...................................................................................................................... 25 Deformation monitoring ............................................................................................................... 27 Built model (SVSolid)..................................................................................................................... 29 Internal erosion factors..................................................................................................................... 30 Material susceptibility................................................................................................................... 30 Hydraulic load ............................................................................................................................... 33 Critical stress condition ................................................................................................................. 36 Failure mode analysis (Critical zones) ............................................................................................... 40 Dam failure........................................................................................................................................ 41 DISCUSSION........................................................................................................................................... 46 Summary of the experimental results .............................................................................................. 46 Summary of the results from the analysis of the internal erosion factors ....................................... 46 Failure Mode Analysis ....................................................................................................................... 47 Dam failure and event tree (Hypothesis verified) ............................................................................ 48 Conclusion ......................................................................................................................................... 48

4

REFERENCES .......................................................................................................................................... 50

Figure 1. Dam, viewpoint from downstream side ................................................................................ 10 Figure 2. Sequence of the experimentation ......................................................................................... 10 Figure 3. Design Model (GBE developed) ............................................................................................. 11 Figure 4. Layering sequences ................................................................................................................ 12 Figure 5. Installation of piezometers in the core .................................................................................. 13 Figure 6. Installation of piezometers top view ..................................................................................... 13 Figure 7. No filter at the crest ............................................................................................................... 13 Figure 8. Vertical pipe ........................................................................................................................... 13 Figure 9. No erosion protection on top ................................................................................................ 13 Figure 10. Grain size distribution .......................................................................................................... 14 Figure 11. Hydraulic conductivity test .................................................................................................. 14 Figure 12. Piezometers ......................................................................................................................... 15 Figure 13. Weir and pressure sensor inside the water tank ................................................................. 15 Figure 14. Turbidity meter .................................................................................................................... 15 Figure 15. Laser scanner 3D .................................................................................................................. 15 Figure 16. Venn Diagram, from (ICOLD, 2013)...................................................................................... 16 Figure 17 Failure mode analysis............................................................................................................ 17 Figure 18. FOS Empty reservoir (upstream) .......................................................................................... 22 Figure 19. FOS empty reservoir (downstream) ..................................................................................... 22 Figure 20. FOS full reservoir (upstream) ............................................................................................... 23 Figure 21. Rapid drawdawn .................................................................................................................. 23 Figure 22. Monitoring results (Core)..................................................................................................... 25 Figure 23. Monitoring results (Filter). ................................................................................................... 25 Figure 24. SVFlux results (Core). ........................................................................................................... 26 Figure 25. SVFlux results (Filter). .......................................................................................................... 26 Figure 26. Deformations monitored downstream (point of view from upstream) .............................. 28 Figure 27. Deformations monitored upstream (point of view from upstream) ................................... 28 Figure 28. Deformation monitored upstream and downstream (point of view from downstream). .. 29 Figure 29. BUILT MODEL-SVSolid deformation results ......................................................................... 29 Figure 30. Comparison results SVSolid, DAM vs BUILT MODEL ............................................................ 30 Figure 31. Grain size distribution .......................................................................................................... 31 Figure 32. Filter Internal stability .......................................................................................................... 32 Figure 33. Core internal stability ........................................................................................................... 32 Figure 34. Pipe internal stability ........................................................................................................... 32 Figure 35. Pore-water pressure (kPa) ................................................................................................... 34 Figure 36. Hydraulic gradients .............................................................................................................. 34 Figure 37. Water head (m) .................................................................................................................... 35 Figure 38. Pressure head (m) ................................................................................................................ 35 Figure 39. Seepage velocity (m/s) ......................................................................................................... 36 Figure 40. Total stress Sy (kPa) .............................................................................................................. 37 Figure 41. XY Displacements (m) .......................................................................................................... 37 Figure 42. Vertical effective stresses (kPa) ........................................................................................... 38 5

Figure 43. XY Shear stresses (kPa) ........................................................................................................ 38 Figure 44. X total stresses (kPa) ............................................................................................................ 39 Figure 45. Y total stresses (kPa) ............................................................................................................ 39 Figure 46. Total minimum principle stress (S3) (kPa) ............................................................................ 40 Figure 47. Local Factor of Safety ........................................................................................................... 40 Figure 48 Critical zones of failure mode analysis ................................................................................. 41 Figure 49. Turbidity ............................................................................................................................... 42 Figure 50. Seepage flow, 10 days (Manual readings) ........................................................................... 42 Figure 51. Seepage flow at day 10 (Thomson weir) .............................................................................. 43 Figure 52. Pore-water pressure at day 10............................................................................................. 43 Figure 53. Acoustic emissions, day 10. ................................................................................................. 44 Figure 54. Dam Failure .......................................................................................................................... 45 Figure 55. Event tree of dam failure ..................................................................................................... 48

6

INTRODUCTION In Sweden, there are two thousand dams providing 65% of the energy demand, most of them are embankment dams. However, some of these 120 large embankment dams are classified as class 1, 2 and 3 (Kraftnät, 2011) .This means that if a failure may occur, these dams would suppose a risk for people´s life and may cause severe economic damages. Even though, failures of embankment dams along the history of this country have not caused any exceptional consequences (Ekström, 2012), there are still uncertainties in the safety of these dams regarding to internal erosion. For instance, the majority of the dams were built between 1950s and 1970s, designed and constructed to standards that might be unacceptable nowadays. Particularly, Vattenfall’s records stated that 40% owned embankments dams developed sinkholes due to internal erosion (Nilsson, 1999) Furthermore, internal threats are difficult to detect and interpret even using new instrumentation techniques. Perhaps, due to the combination of these uncertainties and the state-of-art in the assessment of internal erosion, justify somehow, why approaches in the Swedish dam safety guidelines, RIDAS (2011), are considering conservative assumptions derived from the consequences of this phenomenon, e.g., extreme leakages, instead of assessing the factors of internal erosion within the dam. The results of the assessments with those assumptions usually imply costs of design and construction of new downstream berms, which may not protect the weakest part of the dam against internal erosion (the crest), and may cause associated risks from the new construction. However, why not facing internal erosion firstly, and see if there is any likelihood to occur? In connection to the last question, the aim of this Master Thesis was to face internal erosion before the consequences. The approach was to identify failure mechanisms of embankment dams prior to breaching, and hence verify the reliability of a failure mode analysis after the breaching. Initially, after the literature review, a hypothesis was set as: “If there is internal erosion within the dam, it is possible to analyse the factors that initiate the failure mechanisms, and thus, assessing the safety of the dam prior to breaching”. Later on, in the experimentation phase, the methodology consisted of monitoring a small embankment dam prone to fail by internal erosion mechanisms, whereas, Finite element models (FEM) were implemented to analyse the likelihood of internal erosion prior to breaching. Finally, the hypothesis was verified when the failure mode analysis pointed out the same location where the dam failed in the experiment. This project is the first research project of a program to study the safety of dams funded by Vattenfall Vattenkraft. It took place at Vattenfall Research and Development in Älvkarleby where a new infrastructure for the embankment dam was installed for this purpose.

State-of-art and limitations Internal erosion, by definition, occurs when particles within an embankment or its foundation are carried downstream by seepage flow. It can initiate by concentrated leaks, backward erosion, contact erosion and suffusion (ICOLD, 2013). Even though the problem is known from times of the romans, it was not until the 1970s, when the famous Teton’s dam accident occurred (Solava, 2003), when actually internal erosion was considered a main dam safety issue. This phenomenon has been faced from different perspectives by many authors. Most of the approaches to assess internal erosion have been classified in this literature review in: 1) Hydraulic gradient and grain size of the soils; 2) Dam breach; 3) Surveillance 4) Failure mechanisms and internal erosion factors. 1) Hydraulic gradient and grain size of the soils As Schmertmann (2002) discussed, internal erosion was initially investigated by looking at hydraulic gradients (Bligh, 1910 and Lane, 1935). Later approaches analysed tractive forces of seepage through soil grains (Terzaghi & Peck, 1948 and Sherard & Woodward, 1963). By looking at the particle size distributions of soils it was determined the efficiency of the filters. Sherard (1979) and 7

Terzaghi (1996) made extended experimentation. These investigations developed the filter criteria, which determine if the filters are capable to stop the transport of particles from the core. Meanwhile, experimentation of Skempton (1994) proofed that certain soils may develop internal erosion even at lower hydraulic gradients than Terzaghi critical gradient, gaining more importance the grain size distributions of the soils. Thus, other investigations combined internal stability analysis, such as Kenny and Lau (1985), and probabilistic methods from historical data of accidents. These assessments still were based on the grain size distribution of the core and filters, such as Rönnqvist (2007), Rönnqvist et al., (2014), Bridle, Delgado & Huber (2007), Foster (2001), Brown (2003). One of the problems of these methods in existing dams, older than the method itself, is that the finer part of the soils (sieve size