Вы находитесь на странице: 1из 8

The 12th International Conference of International Association for Computer Methods and Advances in Geomechanics (IACMAG) 1-6 October,

2008 Goa, India

Seepage Analysis through Foundation using F.E.M. and Flownet


Suvasish Mukhopadhyay
Dept. of Civil Engineering, College of Engineering Pune Keywords: Seepage, FEM, Flownet, Permeability, bilinear shape functions ABSTRACT: Seepage analysis in an important factor to be considered in the proper design of any Civil engineering structures. Seepage can occur in both foundations of an engineering structure or sometimes through the structure itself as the case of earthen dams. Seepage analysis pertains to the estimation of quantity of water seeping through the medium in consideration and resulting uplift pressure developed due to it and the exit gradients. Excessive seepage could result in scouring and piping and the ultimate failure of the foundation and the structure accompanying it. Among the various existing methods of analysis of seepage the Finite Element Method and the method of Flow nets are the most widely used method. In the present paper an attempt has been made to combine the two above-mentioned methods. Firstly Finite Element Method is used to determine the nodal potentials of four node quadrilaterals elements. Secondly, flow nets are drawn on the basis of the F.E.M. nodal potentials and bilinear shape functions. In the present paper an attempt has been made to analyze the seepage characteristics below the impermeable structure resting on pervious layer of finite depth underlain by an impervious stratum.

Introduction

Till today many engineers have tried to study the nature of water flowing underground. Various theories and investigations were put forward to predict seepage phenomenon. Still a number of experiments and investigations are in progress to make the engineering structures safe and to utilize this percolating water energy for the benefits of mankind. Various methods are successfully employed to analyze the seepage problems. Sheet piles, grout curtains, impervious cut-off trenches and impervious blankets are the devices widely used to control the seepage. Henry Darcy (1856) proposed a simple relation based on his experiments on les fontaines publiques de la ville de Dijon which is commonly known as Darcys law. Henry Darcy was first to work for finding out the amount of seepage taking place through a homogenous soil. Pavlovsky (1922) gave a general method for determining the functional relationship for confined flow problems, including the closed form solutions of many such problems. The historic work of Dr. Karl Terzaghi (1925) opened the way to scientific and experimental approach to the earthwork design Further the manner in which water particles move through soil pores was put forward by Casagrande (1937). The lucid explanation of practical seepage theory by Dr. Arthur Casagrande (1937, 1961) represents a major s tep in the design of earthwork and other water impounding works. Alam Singh and B.C. Punmia (1973) did the study of uplift pressure below weir aprons. C.S. Desai (1975) developed a numerical procedure based on the finite element method for analysis of steady free surface seepage through foundation & berm of cofferdams. C.S. Desai (1976) has proposed an alternative relaxation type scheme based on residual potential or flow concept for free surface flow through porous media. J.L. Meek and G. Beer (1976) have presented a simple method of plotting a contour through square as well as triangular elements. J.E. Akin and W.H. Gray (1977) have presented a non-linear interpolation procedure for contouring on isoparametric surfaces using standard isoparametric interpolation functions. J.T. Christian (1980) has discussed briefly on the determination of stream functions using the same finite element routines used for solving the potential. J. Aalto (1984) has presented a variational principal and the corresponding finite element equations for the determination of the stream function for soil seepage problems using the standard finite element potential solution as data. J.F. Stelzer and R. Welzel (1987) presented a straightforward scheme for plotting contours that are traced along paths of constant function values. C.S. Desai and B. Baseghi (1988) presented a detailed theoretical development of Residual flow procedure (RFP) for three dimensional seepage, and a scheme for location of the three dimensional free surface. Researcher also verified RPF with respect to (a) closed form solutions, (b)

4175

Variational inequality methods, (c) flow through anisotropic media, and (d) experimental results from glass-bead model. F.T Tracy and N. Radhakrishnan (1989) proposed an analytical method for generating seepage flow lines using the Finite Element Method FEM. This technique includes a new, more analytical (as compared to numerical) approach for determining the boundary conditions for computing the stream functions and a refined approach f or handling the transition between multiple layers. Y. Fan and F.D. Tompkins (1992) presented a simple and unique method for generating flow nets based on nodal potentials and bilinear shape functions. The method reduces the work of performing a second FEM to compute the stream potentials at the nodes.

Method of Design

Problem considered for the proposed investigation is as shown in the fig.1. An impervious structure of width B rests on a pervious soil layer of depth D, underlain by impervious strata. A sheet pile of depthd is located below the top of pervious layer are as shown in figure 1. The coefficients of permeability for the pervious soil in horizontal and vertical directions are kx & ky respectively. The aim is to analyze the effect of the depth of sheet pile d, width of dam B and the influence of anisotropy with respect to permeability of the porous media on the quantity of seepage to which the structure is subjected .
B

B/2

2 H h1

h2

d D F.E. MESH

Where, B = Width of the dam h1 = Upstream water level h2 = Downstream water level H = Total head loss due to seepage through foundation D = Depth of pervious strata d = Depth of sheet pile provided centrally.

Figure.1: Problem considered for proposed

In the present analysis variation of permeability is considered to be parabolic. Following two cases have been studied: a) The permeability at the bottom of pervious layer is taken as unity & permeability at the top of pervious layer is m times that at bottom. In between the variation is considered to be parabolic as shown in fig. 2 (a)

4176

The equation governing this parabolic variation is

x2 =
Where, x = permeability at depth y

(1 m ) y + m
2

y = depth measured from top of pervious strata n = ratio of Permeability at top to that at bottom. d = total depth of pervious strata. b) In this case permeability at the top of pervious strata is taken as unity & permeability at bottom of pervious strata is n times that at the top as shown in fig. 2 (b). Again the variation in between is considered parabolic & the governing equation in this case is

x
Where,

(1 n ) y + n =
2

x = Permeability at depth (d y) d = Total depth of pervious strata y = Distance measured above bottom of pervious layer. m 1.0

1.0 (a) Permeability decreasing with depth

n (b) Permeability increasing with depth Figure 2 Permeability Variation with depth

Five cases with D/B equal to 0.25, 0.50, 1.00, 1.50 and 2.50 have been considered. In each case the sheet pile depth variation is considered as a ratio of sheet pile depth to the total depth d/D, and is equal to 0.8. The upstream and downstream heads have been assigned 30m and 1m respectively. The effect of difference in U/S and D/S heads is not studied in this work and hence the U/S and D/S heads have been kept as constant. under each of the five D/B ratios five sheet pile positions are considered, and for each sheet pile position nine permeability variation cases are considered. Thus, in all two hundred and twenty five cases are considered, and solutions were obtained. The parameter which is considered as constant is d/D = 0.8.

4177

Finite Element Method

In the application of the method of finite element to seepage analyses, the following three fundamental steps are involved. i) Physical Idealization - The purpose of physical idealization is to transform the section of continuous medium under investigation into assembly of discrete size elements, which are interconnected at their nodal points. Within the zones and also within each element the material is homogenous. The element may however possess its own anisotropic character. For the problem under study the idealized assembly has various zones depending upon their D/B ratio, the permeability coefficient for each of them being different. The idealized assembly consists of various numbers of elements interconnected through various nodal points depending upon D/B ratio. ii) Formulation of Equations Potential Function The parent element has reference axes ( ,) with the origin at the center of the square as shown in fig. 3. Assume the flow potential to vary linearly along the edges of quadrilateral element, so as to ensure the continuity of .

1 / 3 i 1

1 / 3 ii i

IV

III

1 / 3

I 1 i

II

1 / 3

ii

Figure 3 Reference axes for Parent Element Thus, by employing isoperimetric characteristics

= N i1 + N ii 2 + N iii 3 + N iv 4 = N p p
i

iv

Where, N p = Linear interpolation function, defined as follows. For two noded parent element = 0 + 1 Considering linear variation 1 = 0 - 1 2 = 0 + 1 Similarly for vertical parent element

= N i1 + N ii 2
Where,

Ni =

(1 )
2

and N ii =

(1 + )
2

4178

Seepage Matrix - Employing the condition that the energy input within the element is equal to the energy dissipated through the seepage. [Se] = F( , ) I + F( , ) II + F( , ) III + F( , )IV

Solution of equations for given boundary conditions For obtaining a non-trivial solution, the equation needs to be modified for given boundary conditions. Two types of boundary conditions exist. a) b) c) d) Stream boundary - The stream boundary is meant to confine the flow and moment a section is idealized; the flow is automatically confined within the section. Potential boundary Equipotential lines - For the plotting of equipotential lines the procedure is that we have all the nodal potentials from F.E.M. The next procedure is to contour the points of equal function values Flow lines - For the plotting of Flow lines method is adopted is similar to the method proposed to Akin and Gray (1977).

iii) Development and Validation of Computer Program - For the proposed investigation, a computer program has been designed in Visual Basic 6. This program performs F.E.M. for the entire mesh and gives outputs into a text file, besides plotting flow lines and equipotential lines. A system compromising of an apron resting over a homogeneous soil stratum equipped with a sheet pile at its center is considered. Here coarse mesh of idealization comprising of quadrilateral elements is employed. 150 quadrilateral elements interconnected through 186 nodes. As far as this problem is concerned following features should be considered: a) b) Due to symmetry around the vertical line passing through the sheet pile, only half of the section needs to be analyzed. In the analysis it is assumed that the lateral extent of 150m beyond the toe of apron represents the indefinite lateral extent of the soil strata adequately.

Values of Nodal Potentials for left hand side mesh and right hand side mesh are calculated iv) Method of Plotting Flow Lines - Flow Lines are plotted by using the following mechanism. The mesh is first prepared and then the each node is assigned the nodal potential value. The global (X, Y) co-ordinates of the nodes are determined. Each element which is surrounded by the four nodes is numbered and their local coordinates are determined. The local co-ordinates are assigned as (s, t), where (-1= s = 1) and (-1= t = 1). The dimensions of the element may vary, but in local co-ordinates the values of s and t never exceed the given range v) Method of Plotting Equipotential Lines - The mechanism is simple and easy to perform; the points having equal function value are connected together. The points are interpolated and connected together. Mechanism for determining the Discharge vi) vii) The difference in U/S and D/S heads is 29m. The numbers of flow lines are 38 and the numbers of equipotential lines are 50. The permeability of the soil is 2m/s. Therefore by using the formulae

Q = Kx H

Nf Nd

Where, H is the head difference between the U/S and D/S ends. H = (301) = 29m. Therefore,

Q = 2 29

Q = 44.08 m 3/s. vii) Solution and Validation of Computer Program

38 50

As the results by hand calculation are available, solution from software is compared with each other. The value for the discharge as obtained by the computer program is 43.55 m3/s.

4179

By comparing the results, for the sample problem considered for checking validity by computer program and by manual solution, it is seen that there is negligible difference between the solutions, which are obtained from manual calculation and from computer programming. Thus it is concluded that the program developed for the proposed investigation is valid for all the cases considered in the investigation.

4 4.1

Results and Discussion Variation of sheet pile position with the discharge for different values of m and n

The Sheet pile position is plotted on the X axis and the discharge is plotted on the Y axis. The discharge variation shows a typical variation with respect to the change in the sheet pile position.
Case m=1
Case n = 2

60 50
Discharge
D/B = 0.25 D/B = 0.5

70 60 50
Discharge
D/B = 0.25 D/B = 0.5 D/B = 1.0 D/B = 1.5

40 30 20 10 0 0 20 40 Sheet Pile Postion 60

D/B = 1.0 D/B = 1.5 D/B = 2.5

40 30 20

D/B = 2.5

10 0 0 20 40 60
Sheet Pile Postion

Figure 4. Variation of sheet pile position with the discharge for various m and n.

4.2

Variation of Depth of Foundation with the discharge for different values of m and n

The depth of Foundation is plotted on the X axis and the discharge is plotted on the Y axis. The discharge variation shows a typical variation with respect to the change in the depth of foundation position.
Case m = 1

Case n = 2

60 50
SP = 0

70 60 50
Discharge
SP = 0 SP = 12.5 SP = 25 SP = 37.5

40
Discharge
SP = 12.5

40 30 20

30 20

SP = 25 SP = 37.5 SP = 50

SP = 50

10 0 0 1
Depth Of Foundation

10 0

Depth Of Foundation

Figure 5. Variation of Depth of Foundation with the discharge for sheet pile position.

4.3

Variation of Nodal Potentials under the apron for different values of m and n

The variations of potentials under the apron are as shown below.

4180

D/B = 0.25, Sheet Pile Postion = Upstream End 45 40 35 Nodal Potentials 30 25 20 15 10 5 0 0 0.2 0.4 0.6 0.8 1 1.2 Fractional Distance Form Upstream Side
m =1 m=2 m=4 m = 10 m = 100 n =2 n=4 n = 10 n = 100

Figure 6. Variation of Nodal Potential with the discharge for various m and n. With D/B ratio 0.25, 0.5, 1.0, 1.5 and 2.5 graphs are plotted for various locations of sheet pile e.g. sheet pile at upstream end, B/4, B/2 and 3B/4 for different values of m and n. By these way we get 20 sets of family of curves. From the studies of seepage characteristics made for the impervious structure the following conclusions are made. 1. 2. 3. For any permeability variation and for same value of m or n the potential values at any point on floor base increases as the D/B ratio increases. For same values of D/B and m or n the potential values decrease from upstream to downstream along the floor base. As the position of sheet pile varies for a specific D/B ratio from the upstream to the downstream end of the apron there is initially an increase of discharge until the sheet pile position is at th position and then there is a sudden decrease in the discharge. As the D/B ratio increases there is a decrease in the discharge of water seeping under the apron. As the permeability increases for a specific sheet pile position and a fixed value of D/B ratio the dis charge increases. As the permeability decreases for a specific sheet pile position and a fixed value of D/B ratio the discharge increases.

4. 5. 6.

References

Alam Singh and B.C. Punmia, Seepage below horizontal apron with downstream cutoff, founded on anisotropic pervious medium of finite depth, Indian Geotechnical Journal , vol. 3, Number 3, July 1973, pp 198 - 215. C.S. Desai, Free surface seepage through foundation and berm of cofferdams, Indian Geotechnical Journal, vol. 5, Number 1, January 1975, pp 1 10. C.S Desai, Finite element residual schemes for unconfined flow, International Journal for numerical methods in engineering, 10, pp 1415-1418. (1976). C.S. Desai, Finite element methods for flow in porous media, in Gallagher, et al. (eds.), Finite elements in fluids, Chap. 8, Wiley, Chichester, 1975, pp 157- 182. C.S. Desai, Elementary finite element method, Prentice-Hall, Englewood Cliffs, New Jersey, (1979). C.S. Desai and B. Beseghi, Theory of residual flow procedure for 3-D free surface seepage, Advance in water resources, Vol 11, No 4, pp192 203. (1988). D.V. Griffith and I.M. Smith, Programming the Finite Element Method, IIIrd Edition, John Wiley and Sons, West Sussex, England (1996). F.T. Tracy and N. Radhakrishnan, Automatic generation of seepage flow nets by finite element method, Journal of Computing in civil engineering, Vol 3, No 3. July (1989), pp 268 284. J. Aalto, Finite element seepage flow nets, International Journal for numerical and analytical methods in Geomechanics , 8, pp 297 303. (1984). J.E. Akin and W.H. Grey, Contouring on isoparametric surfaces, International Journal for numerical methods in engineering, 11, pp 1893 1897. (1977). J.F. Stelzer and R. Welzel, Plotting of contours in a natural way, International Journal for numerical methods in engineering, 24, pp 1757 1769. (1987).

4181

J.L. Meek and G. Beer, Contour plotting using isoparamertic element representation, International Journal for numerical methods in engineering, 10, pp 954 957. (1976). J.T. Christian, Flow nets from finite element data, International Journal for numerical and analytical methods in Geomechanics , 4(2), pp 191 - 196. (1980). K.A. Siraskar and Dr. V.J. Patel, Equivalent permeability of non homogeneous soils, Journal of Indian National Society of Soil Mechanics and Foundation Engg, Vol.No6, April (1967), pp 153 - 165. Y.Fan and F.D. Tompkins et al, Generation of flow nets using F.E.M. nodal potentials and bilinear shape functions, International Journal for numerical and analytical methods in Geomechanics , Vol 16, pp 425 437. (1992). Zienkiwicz O.C., The Finite Element Method in Engineering, McGraw Hill Book Co. London, 1971.

4182

Вам также может понравиться