Effect of Number of Layers on Incremental Construction Analysis of Earth and Rock Fill Dams

 

Ali Zomorodian

E-mail: smazomorod@yahoo.com or mzomorod@Shirazu.ac.ir

K. Sahebzadeh

and

A. Ooria

Department Of Irrigation, University of Shiraz, Shiraz, Iran

 

TECHNICAL NOTE

 

ABSTRACT

The construction period and the end of construction have an important effect on the design and the performance of earth and rock fill dams. It has been known pore water pressure developed during this time in the impervious section is an important factor in slope stability analysis of dams. In this paper the parameters, which their values significantly change with selecting different numbers of layers in computer model simulation of dam, are investigated. Sange Siah earth dam is analyzed with different numbers of layers and then results are compared with several other real cases. From these results one can propose that the minimum number of layers, which must be included in the analysis of earth dam with vertical clay core, is equal to ten. The slope stability analysis in this research is done with Geo-slope software.

Keywords: stress, strain, finite element, consolidation, Sange Siah dam

INTRODUCTION

As a part of the design process of earth or rockfill dams the engineer must estimate the movements, strains, and pore water pressures which may develop in the dam at various times in its life. Under certain conditions during and at the end of construction, the embankment compression due to weight of the overlying fill causes high saturation degree of clay core materials and there is not enough time to dissipate pore water pressure. Therefore, high pore water pressure may develop in the impervious sections of earth dams and it may control the design of the dam from the standpoint of slope stability, deformation, and hydraulic fracture. The prediction and the control of pore water pressure have a significant effect on the performance of dam.

Since numerical methods especially the finite element has a capability to analysis nonlinear stress-strain including complicated boundary conditions, irregular geometry, and some different conditions can implement. The study of the observed deformations, stress and pore pressure distribution and finally interpretation of the result led to evaluate true dam behavior. In this paper the effect of numbers of construction layers in computer model simulation of dam with considering time, drainage direction, and decreasing pore water pressure (due to consolidation) are investigated. At the end results from the point of view of different parameters are compared. To investigate the effect of numbers of construction layers a case study is done on Sange Siah dam, which is an earth dam with vertical clay core.

The analysis is done by Geo-slope software, which is based on the finite element. This program is capable of doing slope stability analysis by limit equilibrium. And doing total and effective stress also consolidation analysis and stage construction model. The principle is used in the Geo-slope software is based on the unsaturated soil mechanics (Fredlund) and geotechnical earthquake engineering ( Karmer).

STUDY OF EMBANKMENT ANALYSIS WITH DIFFERENT NUMBER OF LAYERS

Earth and rockfill dams usually raise about 15 cm thickness per day. Therefore numerical modeling of stage construction of an embankment requires a thicker layer. The number of layers which is required depends on the data and capability, precision and computation time of computer program. To study foundation of embankment only a few number of layers is required. Since the stress and displacement are not sensitive to numbers of layer. Therefore, the assumption that the embankment is constructed in one layer is sufficient. But for embankment more layers is necessary. To compute stresses several layers or even one layer and for displacement more layers in needed. In the analysis of large earth dam usually ten layers are used. This number of layers can be reduced without effect on the results. Naylor and Jones 1973 are shown for the embankments that its width in comparison with height are large (1 D) using five layers gives good approximation as comparing with infinite layers of soil with high nonlinear behavior. There is not any agreement between predicted displacements in one layer model with actual displacements. The maximum predicted settlement for one layer model occurred at the crest but actual maximum settlement is about center of the embankment [1]. It follows from the fact that the lower portion of the dam consolidates partly or completely, with the development of respective settlements, while the upper portion is still under construction [2]. Vertical stress components at the embankment can compare with stress from analysis of several stages construction [1].

Kinda dam is a rockfill dam with vertical clay core built on incompressible rock. Its height is about 72.6 m and maximum width at the bottom is 315 m and clay core width is 54 m. The upstream slope is combination of 2:1.5 and 1:1.8 and downstream slope is 1:1.9. Material type and range of plasticity used for Kinda clay core are similar to Sange Siah dam. The finite element simulation of Kinda dam was done in 9 steps which height various between 5.5 to 9.1 m [2].

 


Figure 1. Rockfill dam (kinda) with vertical earth core on incompressible rock. Development of core settlements due to load [2].

 


Figure 2. Computed lines of equal vertical settlements (cm) at end of construction [2].

Regarding to the results of settlement analysis and observed settlement for this model it can extend the result reported by Wheeless & Sowers (1972) that in all cases, the estimated settlement exceeded the observe settlement and 2/3 up to 4/5 of total observed settlement was completed during the period of embankment construction.

SANGE SIAH DAM AND ITS MATERIAL PROPERTIES

Sange Siah dam is an earth dam with vertical clay core. It is constructed on Jam Sange Siah river located in Kurdistan of Iran. A maximum height of 33 m from bedrock and crest length of 351.6 m and the crest width is 10 m. The maximum width at bottom is 180.5m. The upstream and downstream slope is 1:3 and 1:2.5 respectively. The clay core has a slope of 1:0.25 and maximum width of core is 19.5 m.

Material properties of Sange Siah dam are given in Table 1. Linear elastic model is used to analyze the dam behavior. Geo-slope software has capability to introduce material properties close to field condition and unlike other software enable to analyze unsaturated medium. The finite element mesh of dam with thickness of 1 m is shown in Fig. 3.

Table 1. Material properties of different zones of Sange Siah dam

 


Figure 3. Finite Element Mesh of Sange Siah Dam

To ensure high precision this small thickness is selected. In this study 10 models with different numbers of layers are analyzed. The numbers of layers are 1, 2, 3, 4, 5, 6, 8, 10, 15, and 31. The total construction time 186 days is selected. In each stage time for construction of each layer to allow drainage is define. For example in a model with 31 layers time for construction each layer is 6 days. The maximum values of some parameters, which control the analysis of above models, are shown in Table 2. If the analysis of 31 layers model, which has the minimum thickness for each layer and has close result is compared with computed model and observed measurement of Kinda dam considering as a criteria.

Table 2. Maximum Values of some parameter from analysis of tenth models of Sange Siah Dam
Steps Vertical Deformation M Total Stress (X) (kPa) Total Stress (Y) (kPa) Effective Stress (X) (kPa) Effective Stress (Y) (kPa) Pore Water Pressure (kPa) Shear Stress (X-Y) (kPa)
10.515 355 621 311 577 116 64
20.442 363 634 316 588 67 69
30.462 360 635 321 595 55 68
40.439 360 636 322 599 52 67
50.422 359 634 321 597 50 67
60.404 358 634 321 596 49 67
80.397 357 632 321 596 48 66
100.380 357 630 319 593 47 66
150.367 356 628 318 590 47 65
310.348 360.4 626 309 574 51 64

The difference between total, effective and shear stresses of these models are negligible. But difference in vertical settlement and pore water pressure is significant. The percentage difference of pore water pressure for 31 layers model compare to one, two and 3 layers model are 127%, 31% and 7% respectively and for other models the difference is negligible. As the number of construction layers increases percentage difference of vertical settlement from 1 to 31 layers model decreases and are equal to 48%, 27%, 23%, 33%, 26%, 21%, 16%, 14%, 9.2%, 5.5% respectively. But it must be noted these values show distribution of parameters in different part of dam. Figure 4 shows settlement of central core at different level during construction at 0.25 H, 0.5H, 0.75 H and end of construction for selecting 1, 2, 3, 6, 8, 15, 31 layers in a model. Construction stages at level of 0.25 H is completed after 48 days, level of 0.5 H after 96 days, level of 0.75 H after 138 days and end of construction after 186 days. The result of these tenth models in form of contour line can be shown. Vertical settlement for 1, 3, 5, 8, 15, 31 layers models are shown in Fig. 5. From these contour line one can observe that definition one or several layers model has not significant effect on distribution of all mention parameters except vertical settlement.

 


Figure 4. Development of core settlements in Sange Siah dam at different level during construction at 0.25 H, 0.5H, 0.75 H and end of construction for selecting different number of layers.

 



Figure 5. Settlement distribution for a model with different number of layers
(a) 1, (b) 3, (c) 5, (d) 8, (e) 15, (f) 31.

 

CONCLUSION

The result of pore water pressure for dam analysis with one layer model is conservative and is not useful for true evaluation or even economical evaluation of dam. In general if the objective of study is stresses one can use single layer in analysis. From diagrams of dam settlement verses height for analyzing with 2, and more layers is observed that approximately concentrated settlement in each stage located at mid thickness of that layer. Comparing Fig.2 for 2 and more layers models with Fig.1 which shown for Kinda dam, it is observed that concentrated settlement located at 0.5 – 0.7 H ( H = dam height). Since it follows from the fact that the lower portion of the dam consolidates partly or completely, with the development of respective settlements, while the upper portion is still under construction and with increases number of layers in models concentrated settlement moved to mid height of dam. Comparing estimated and observed the result of Kinda and Masjed Soliman dam (in Iran with 177 m height and using 17 layers in modeling of dam [4] ) an earth dams with vertical core with analysis of Sange Siah dam one can recommended that at least number of layer for reasonable results at numerical model is equal 1/9-1/10 H. It must be mentioned that this study is done only for dams with vertical clay core without considering foundation in analysis (competent foundation).

REFERENCES

  1. Naylor, D. J. and G.N. Pond (1981) “Finite Element Method in Geotechnical Engineering” Mir publishers.

  2. Kutzner, C. (1997) “Earth and Rockfil Dams,” published by A. Balkema, Rotterdam/ Brookfield.

  3. Das, M. Braja (1990) “Principles of Geotechnical Engineering” PWS-KENT Publishing Company, Boston, U.S.A.

  4. Jafarzadeh, F. and M. Talebi (2002) “Evaluation of Rockfill Dam with Unsaturated Clay Core During Construction” Third Iranian International Conference on Geotechnical Engineering & Soil Mechanics Tehran, Vol. 2, pp. 308-314.

  5. Wheeless, L. D. and G. F. Sowers (1972) “Mat Foundation and Preload Fill, Va Hospital, Tampa,” Proceedings, Specialty Conference on Performance of Earth and Earth- Supported Structures, ASCE, Vol. 1, Part 2, pp. 939-951.

    6.  

    © 2006 ejge