Physical model in hydrogeology

sand tank model

Arab Sanaa, Madene sawab, Medjdoub amel, Moulay lakhdar sohaila

Abstract

Keywords: sand tank model, hydrogeology, physical model, aquifer, well

1. Introduction

Groundwater is one of the hydrological system's main pillars (especially in terms of clean water supply).
Technically it’s the Water stocked in the zone of saturation (The part of the soil mantle and permeable
bedrock where water fills all of the pores). A geologic formation capable of storing, transmitting and
yielding groundwater to wells is called an aquifer. Groundwater in the zone of saturation comprises most
of the globe’s usable fresh water supply. An aquifer that has no confining layer between the land surface
and the saturated zone is an unconfined aquifer. The top of the saturated zone in an unconfined aquifer is
the water table. An aquifer sandwiched between impervious or slowly permeable materials is a confined
aquifer. Confined aquifers may be exposed to the surface at some point. That part of the aquifer exposed to
the surface would then be unconfined. The imaginary upper surface caused by hydraulic pressure in the
confined aquifer is the potentiometric surface. The potentiometric surface is the level to which water
would rise in an unpumped well put into the confined aquifer. A confined aquifer is sometimes called an
artesian aquifer. An artesian well is a well into a confined aquifer. A flowing artesian well occurs if the
potentiometric surface of the aquifer is above the ground surface.

Unfortunately, both the quality and quantity of these waters represent a major concern that’s why science
is working constantly to fin d some effective tracking tools to expect and prevent the problems that can
occur within these waters, among the proposed solution the models stand as very accurate and efficient.

In groundwater studies, models are developed and applied to predict the fate and movement of
groundwater physiochemical aspects in natural as well as hypothetical scenarios. these studies allow us to
transform the data to a set of information which helps building the needed knowledge to take and some
asses a decision, it serves as an investigation material in the hydrodynamic studies. Groundwater models
are designed to be a set of simplified versions of the real groundwater scenarios. They are used to
simulate and predict aquifer conditions. Sand tank models, analogue models and mathematical models are
the broad categories of groundwater models.

And in this article we will try to focus on the sand tank model. A tool that proved itself to be very effective
when it comes to facilitating the visualisation of the ground water movement. if described in the simplest
way we can say that sand tank groundwater model is a glorified fish tank full of unconsolated porous sand,
equipped with a discharging and charging in/outlets allowing the water to go through it (an induced
flow) just like an aquifer and unlike other models it can extend to a large scale of uses. as it can
demonstrate the water cycle, the ground water flow, confined and unconfined aquifers, wells and their
pumping effect, contamination and sea water intrusion...nevertheless we have to mention that the
phenomena measured at the scale of a sand tank are often different from conditions observed in the field
and conclusions drawn from this kind of models may need to be qualified when translated into a field
situation (we will emphasize further later on).

The goal of modelling is to predict the value of an unknown variable (for example we have the ground
water head or the concentration of the contaminant) at nodel points. they are often used to the effects of
pumping on ground water levels. as for this following article we will use the sand tank model in two of its
main application “the study of flow analysis in semi confined aquifers “and “the study of horizontal
laminar flow” starting from their theorical (maths and physics) parts to results uses advantages and limits.
2. Physical models

Physical model (also called scale models) is actual physical replicas of a groundwater flow system (mostly
a simple aquifer) that have been scaled down for study in the laboratory. For example, the soil column
(one-dimensional) and the sand tank (three-dimensional) are physical models. The behavior of the
prototype to hydrologic stresses can be simulated by subjecting the scale model to certain stresses such as
water removal or injection or contaminated recharge. Scale relationships are used to interconnect the
prototype with the physical model, for example Uξ=ξm/ξ, where U is the ratio of the model parameter ξm
divided by the actual aquifer parameter ξ.

3. Definition of sand tank model

One of the best tools we have is what we call the Sand Tank Groundwater Model. It’s a small Plexiglas
container that represents a cross section of the earth’s crust. We can add water to it, we can add dye to it
which helps visualize how that groundwater moves through the landscape to maybe a lake, river or
stream and can illustrate how things we do at the landscape can impact the quality below. It’s also dye to
simulate pollutants, or pumping wells to simulate groundwater withdrawals; viewers are better able to
visualize and understand complex groundwater principles.

The sand tank is a true model in that the same laws governing the flow of water apply to both the model
and aquifer. Sand tank models are usually constructed in radial sector, columnar, or rectangular shaped
containers filled with a porous medium. The containers are most often made with transparent glass or
plastic walls so that the flow can be observed. Porous media used in sand tank models range from
commercially bagged silica sand, crushed sandstone, and graded river sand, to uniform-sized glass
spheres. The porous material is packed under water to achieve a homogeneous and isotropic media.
Piezometers are installed through the model walls at points of interest. Suitable pumps and flow rate
measuring devices are connected to the model to control the flow of fluid. Fluids used in the sand tank
model for studying single phase flow problems have been ordinary tap water, ethanol-water mixes, and
mineral oil. There are so many problems that are resolved by sand tank models which are mentioned
below:

- groundwater re- charge mechanisms (Marmion, 1962),

- dispersion phenomena (Bear, 1961);
- density induced mixing (Gelhar et al., 1972);
- and salt-water intrusion;
4. Some applications of sand tank model
- Sand tank can be used to demonstrate hydraulic gradients or well interference
- It can be used also as processes like groundwater–surface water interaction
- Then like contaminant (food dye) transport.

These physical models allow students to visualize subsurface processes—and then change the forcing
functions that affect them—providing a hands-on way to help students intuit behavior in field settings.

In addition, sand tank model has some types which vary according to aquifer features
 Type of model Purpose of model

Vertical cross section , 2D,flow non steady Study perched water tables

3D flow ,90° radial sector, steady Clarification of the nature of unconfined flow

Vertical tube ,1D flow, steady Study of dispersion phenomena 2D and 3D

Christiansen filter, 2D non steady

Vertical cross section, Christiansen filter, 2D Investigation of groundwater mounds due to

non steady artificial recharge

Vertical cross section, 2D flow non steady Study free-surface flow to outflow channels

Vertical cross section,2D flow steady Investigation of gravitational convection and

dispersion of salt water
TABLE-0

SOME APPLOCATION OF SAND TANK MODEL

5. Advantage of sand tank model

The advantage of sand tank is that it can be easily emptied and washed clean of all earth materials that
have been previously loaded. This enables the researcher to change the soil profiles to fit the geographic
area that is being investigated.

A particular advantage of the sand tank model is that it is well suited for studying the effects of dispersion,
diffusion and other multiphase flow problems since the microscopic structure of the sand tank model
medium is of the same nature as that of an actual aquifer. However, according to Bachmat same nature as
the severe restrictions on choice of scale factors in studying dispersion phenomena

The definite advantage associated with the visual natural nature of a sand tank model. flow lines can be
made visible by dye injection. Bear (1961) points out how tracer concentration charges can be detected by
photoelectric cells which sense variation in transmitted light through a model

6. The disadvantage of sand tank model

A disadvantage of the sand tank model is that regional flow problems involving aquifers of large
horizontal areal extent have not been studied.

This is probably due to the restrictions imposed by the vertical size model, the vertical depth would be
small, thus giving rise to a very small time scale which is very difficult to instrumentate under artesian
condition, and problems are compounded

Kraijenhoff Van De Leur indicates that the capillary rise in sand tank models, operate under unconfined
condition, is larger than that occurring in an actual aquifer situation. He elaborates on special scale factors
which must be used to bring the unsaturated flow contribution under proper scale

7. STUDY OF FLOW ANALYSIS IN SEMI-CONFINED AQUIFER USING A SAND TANK MODEL

7.1 Theoretical Analysis

The theoretical approach to the problem of flow of water to partially penetrating well in leaky aquifer
based on basic flow theories is presented here by using dimensionless approach to obtain general solution
of the flow problem. The flow in partially penetrating well in a leaky aquifer depends on the following
factors: (I) Discharge from the well, Q P; (2) Hydraulic conductivity of the aquifer, K; (3) Hydraulic
conductivity of the leaky aquifer, K'; (4) Radius of influence, R; (5) Radius of the well, r (6) Drawdown, s;
(7) Thickness of the aquifer, B; (8) Thickness of the leaky aquifer, B'; (9) Depth of penetration of the well,
I, where I= pB, p =Degree of penetration; (I 0) Operation head, H.

More conveniently the discharge can be written as a function of other variables as follows:
QP = f1 (K,K’, B,B' ,R,r,s,l,H) (I)

Using standard methods of dimensionless analysis, the equation can be written as,
QP /KH 2 =f2 (s/H , l/H, R/H, r/H , B/H , B'/H) (2)

The dimensionless terms r/H, R/H, B/H, B'/H are constants throughout the experiment.
Hence, they can be eliminated by considering them as constants. Thus, Eq. (2) can be rewritten
as follows:
QP /KH 2 =f3 (s/H , l/H) (3)

For a fully penetrating well the discharge depends on every other factor as mentioned

above except 'I'. Here the term 'p' becomes equal to one. Hence, by the same methodology,

Q = cp 1 (s,B,B', K,K' ,R,r,H) (4)

or Q/KH2 =cp 2(s/H, B/H, B'/H, R/H, r/H) (5)

or Q/KH2 =cp 2(s/H) (6)

From Eqs. (3) and (6) the discharge ratio can be written as:

QP/Q=cp(s/H,l/H) (7)

7.2 Assumptions

The assumptions considered in solving the flow problem are:

- The aquifer is isotropic, homogeneous and of infinite radial extent.

- Aquifer is pumped at constant discharge rate.
- The phreatic surface remains essentially constant.
- The flow is radically symmetric and is laminar.
- Flow through the bottom of the aquifer is negligible.
- Dupuit's assumptions are valid

7.3 Components of Model:

The model consists of a rigid foundation which serves twofold purposes one that checks the percolation
the percolation loss through the bottom and the other that gives rigidity to the model. Piezometers were
provided to measure the static heads at different points. They are provided in three rows (Fig. I), each row
having ten numbers of piezometers with 0.635 cm external diameter. The center to center distance of
piezometer tubes are 23 cm, and these are connected to glass tube fixed on a manometer board with the
help of P.V.C. tubes. Sump meant for supplying water at constant head was located at a distance of 245 cm
away from the center of the test well. The maximum level of water in the sump was 107 cm from the
bottom. To maintain constant head, water was recirculated and also supplied from outside to make up any
unavoidable loss. The test well was a slotted pipe of II cm diameter having a slotted strainer extended
throughout the depth of the aquifer. An arc of 45 degrees is actually exposed to the aquifer and the other
slots are closed as these are connected to the model well. The bottom of the well was closed to prevent
entry of water through it. A screen made out of 40-mesh wire netting was used to separate the sump and
the aquifer materials. It allows the water to pass through it and prevents the aquifer materials from falling
into the sump. A Tulu pump (H.P=0.025, Watt=40, rpm=3000, Suction head=0.61 m, Delivery head=4.6 m)
was used to deliver water.
 FIG.1 SAND-TANK MODEL

7.4 Selection of Aquifer Materials:

Capillary rise in sand tank model is disproportionately large compared to that occurring under field
conditions. However, in leaky-confined aquifer these effects are more or less insignificant. But in
unconfined aquifers these corrections are necessary. For this experiment sand taken for aquifer material
was relatively course. The sand taken for main aquifer was evenly mixed, homogeneous and had natural
texture. For making leaky zone sand was mixed with silt in a proportion of 3: I (by volume). The material
used for the main confined aquifer that is sand for the experiment had the average particle size of 0.43
mm and uniformity coefficient of 2.26. The material used for leaky zone i.e. sand-silt mix had an average
particle size of 0.26 mm and uniformity coefficient of 2.21. In this experimental set-up an artificial gravel
pack was used near the screen openings as per U.S.B.R. criterion.

7.5 Determination of Hydraulic Conductivity:

Hydraulic conductivity can be determined in two ways for this experiment. Firstly, it can be determined
with the help of permeameter where the disturbed sample of the aquifer material is used. Secondly, since
the maximum drawdown in fully penetrating well in leaky-confined aquifer is known, the readings can be
used in deGlee's equation to get the values of hydraulic conductivity in actual model environment.

7.6 Determination of Drawdown:

Piezometers are meant for giving direct of drawdowns at different radial distances for each discharge. In
this experiment the sump was filled to maximum level of 107 cm and the aquifer was allowed to get
saturated. The fall in water level in piezometer are used to measure drawdown for different points.
7.7 RESULTS AND DISCUSSIONS

Experimental result on physical properties of aquifer material such as hydraulic conductivity, the
relationship between discharge and drawdown in fully penetrating well and effect of partial penetration
on discharge in leaky-confined aquifer are discussed here. Hydraulic conductivity by constant head
penneameter by using Darcy's law for sand and sand-silt mix was found to be 9.720 and 1.0793 m/day.
Since the dependability of the results obtained from laboratory method is less, the data obtained has been
used in deGiee's equation and the hydraulic conductivity was found to be 7.814 m/day for sand and 0.225
m/day for sand-silt mix. The drawdowns corresponding to various discharges and penetration ratios are
presented in Table I.

TABLE-1

OBSERVATIONS OF DISCHARGES AT DIFFERENT DRAWDOWNS

SI. Drawdown Discharge for different penetration rations (cc/s)

NO. (cm) 100% 80% 60% 40% 20%
1 5 4.2 3.77 3.37 3.37 2.78
2 10 8.27 7.12 6.33 5.6 4.13
3 20 15.7 14.2 13.25 11.30 8
4 30 24.53 21.87 19.43 15.78 10.7
5 40 35.33 32.1 28.17 20.4 -
 Discharge rate represent for a maximum drawdown of 27.1 cm.

TABLE-2

OBSERVATIONS OF DRAWDOWNS AT DIFFERENT RADIAL DISTANCES

FOR FULLY PENETRATING WELL

SI. Drawdown Drawdowns at different radial distances in cm

NO. (cc/s) 11.5 34.5 57.5 80.5 126.5 218.5
1 35.33 40 22.5 15.1 10.4 6.1 0
2 24.53 30 15.6 11.1 6.5 3.9 0
3 15.7 20 9.8 8.1 4.5 2.3 0
4 8.27 10 4.7 3.6 2.1 0.6 0
5 4.2 5 1.5 1.1 0.8 0.2 0

It is evident from Fig. 2 that to get a particular value of discharge from the well, the drawdowns are
different for different penetration ratios. Drawdowns required, for getting a certain discharge, goes on
increasing as the penetration ratio decrease. Further, the incremental increase in draw down for getting
the same discharge increases as the penetration ratio decreases. At lower penetration value of 20 % and
40 %, the variation of discharge with drawdown is very small thus giving a smaller value of maximum
drawdown. It is observed that at higher penetration the curve is not straight, rather it is slanting
downwards in the upper part. From the results given in Table-1 it is observed that at higher drawdowns
when the piezometric pressure (between phreatic surface and confined aquifer) is high, the contribution
to flow from the overlying layer through leaky zone seems to be significant. It is more prominent near the
vicinity of the well where there is a sharp drop in piezometric head. With decrease in penetration ratio,
the vertical flow through the leaky zone though still present, gets mixed up with the horizontal flow and
the resultant takes a path towards tube well face. But the major contribution to the yield at lower
penetration ought to come from the horizontal component.
 FIG. 2 DISCHARGE-DRAWDOWN RELATIONSHIPS

The data presented in Table 2 was plotted on semi-log graph depicting radial distance drawdown
relationships for a given discharge. The curves as shown in Fig. 3 are bent upwards and are not straight
lines as was observed by deGiee. However, these curves become straight t lower discharge values or when
the steady state drawdown is lowered. As may be recalled, for steady state flow in confined aquifers
according to Theim's equation, drawdown is proportional to 'log r' for a given discharge. Therefore, the
drawdown radial distance plot on a semi-log graph is a straight line. But in case of a leaky-confined aquifer
yielding steady state flow, maximum drawdown (Sm) at different observation well points are a function of
'Q' and 'K0 (r)' according to deGlee. And according to Hantush and Jacob 'Sm' is a function of 'Q' and 'log r'.
Both of these equations on a semi-log plot will give a straight line. However, the curved shape of the 's-log
r' plot (Fig. 3) shows that discharge varies along the radial line. This is due to the fact that; the leaky-
confined aquifer gets vertical flow from overlying zone through leaky layer increasing the discharge
incrementally. Unlike confined aquifer where the vertical flow is absent, the discharge no longer remains
constant but goes on increasing, as one moves towards the well. Also from the model experiment it is
observed that the curve falls steeply near vicinity of the well, where there is a sharp increase in discharge.
It is also observed that, for lower values of maximum drawdowns, the profile of the curve assumes more
or less a straight line at low piezometric head (between phreatic surface and confined aquifer) where the
contribution of vertical flow from overlying zone is insignificant. Thus the discharge at any point radially
from the well is maintained more or less constant for lower values of maximum drawdowns.
 FIG. 3 DRAWDOWN-RADIAL DISTANCE RELATIONSHIPS FOR A FULLY PENETRATING WELL

A non-linear relationship was tried to be established between drawdown, discharge, penetration ratio and
radial distance. The results are given in Table 3, 4 and 5. The results indicated that the parameter 'A' in the
relations= A (log r) b drastically decrease in Q, thus for a given higher discharge value the decrease in
drawdown is sharp with radial distance making the s-log(r) plot nonlinear. In the relationship
Q=A. (s) b (P )c, penetration ratio parameter 'c' increases with increase in drawdown. It indicates clearly
that the discharge (Q) increases sharply with the drawdown compared to corresponding increase in
degree of penetration. Further regression analysis was carried out on experimental data with s and p as
independent variable and Q as the dependent variable. The results of regression are shown in form
relationship in following equation.

Q=0.68 s+ 0.01 p- 6.74 (8)

TABLE-3

RELATIONSHIP BETWEEN SAND LOG 'R' : S=A (LOG R)b

SI Coefficient of parameter Discharge(Q)

correlation (R2) A B In cc/S
1 0.918 47.316 -2.322 35.33
2 0.970 35.606 -2.465 24.53
3 0.968 23.678 -2.462 15.7
4 0.933 11.974 -2.651 8.27
5 0.984 6.129 -3.03 4.2
 TABLE-4

RELATIONSHIP BETWEEN Q, SAND LOG P: Q = A.(S)b (P)c

SI Drawdown parameter Discharge(Q)

In cm A B C In cc/S
1 5 2.966 0.192 0.269 0.929
2 10 3.927 0.812 0.421 0.986
3 20 1.205 0.851 0.370 0.976
4 30 4.249 0.517 0.501 0.997
5 40 2.796 0.693 0.567 0.975

TABLE-S

RELATIONSHIP BETWEEN Q, SAND LOG R : Q = A.(S)b(LOG R)c

SI Radial distance parameter Coefficient of

In cm A B C Correlation (R2)
1 11.5 0.622 1.096 -0.358 0.997
2 34.5 1.875 0.922 0.125 0.996
3 57.5 2.048 1.088 -0.224 0.984
4 80.5 4.162 0.896 0.075 0.995
5 126.5 0.624 0.702 3.694 0.998

8. Conclusion:
Reference

- STUDY OF FLOW ANALYSIS IN SEMI-CONFINED AQUIFER USING A SAND TANK MODEL P. K. Bhunya a &
B. Anjaneyulu ba National Institute of Hydrology, Roorkee b IIT , Kharagpur Published online: 07 Jun 2012.

-Linking Physical and Numerical Modelling in Hydrogeology using Sand Tank Experiments and COMSOL
Multiphysics, Kamini Singhaa* and Steven P. Loheide IIb: 28 July 2010

Liste of figures

FIG.1 SAND-TANK MODEL

FIG. 2 DISCHARGE-DRAWDOWN RELATIONSHIPS

FIG. 3 DRAWDOWN-RADIAL DISTANCE RELATIONSHIPS FOR A FULLY PENETRATING WELL

Liste of tables

TABLE-0 SOME APPLOCATION OF SAND TANK MODEL

TABLE-1 OBSERVATIONS OF DISCHARGES AT DIFFERENT DRAWDOWNS

TABLE-2 OBSERVATIONS OF DRAWDOWNS AT DIFFERENT RADIAL DISTANCES FOR FULLY

PENETRATING WELL

TABLE-3 RELATIONSHIP BETWEEN SAND LOG 'R’: S=A (LOG R)b

TABLE-4 RELATIONSHIP BETWEEN Q, SAND LOG P: Q = A.(S)b (P)c

TABLE-S RELATIONSHIP BETWEEN Q, SAND LOG R: Q = A.(S)b(LOG R)c