- Email: [email protected]
Utilization of local material in the construction of an embankment for recharging groundwater aquifer with treated wastewater effluents
8)
Pergamon
War. Sci. Tech. Vol. 40, No.7, pp. 25-32.1999 C 1999IAWQ Published by Elsevier Science LId
Prinled in Oreal Brilain. All rights reserved
0273-1223J99 $20.00 + 0.00
PD: 50273-1223(99)00580-6
UTILIZATION OF LOCAL MATERIAL IN THE CONSTRUCTION OF AN EMBANKMENT FOR RECHARGING GROUNDWATER AQUIFER WITH TREATED WASTEWATER EFFLUENTS S. A. Aiban, A. M. Ishaq and M. S. AI-Suwaiyan Department of Civil Engineering, King Fahd University ofPetroleum &< Minerals. Dhahran 31261. Saudi Arabia
ABSTRACT Water is certainly the moSt precious and valuable resowce of the physical environment for all living creatures. It is now well established that in many agricultural regions in the Kingdom of Saudi Arabia. the water table has been depleted dramatically. If the current level of agricultural production is 10 be maintained. other sources of agricultural water must be found. There is at least a billion cubic metres of secondary municipal effluent wasted annually. Thus. utilization of these wastewater effluents to recharge groundwater aquifers and reuse them at least for agricultural purposes becomes a viable proposition. In the study reported herein, treated wastewater will be used 10 recharge aquifers using spreading basins. The basin consists of a dune sand filter confined. from all sides. by an embankment of compacted sand/marl material. A field site has been constructed in AJ-Aziziyah in eastern Saudi Arabia. The sand filter is constructed from clean dune sand. has an area of 25 x 25 rn, and has an effective height of approximately 7.0 m. The stability and permeability of the embankment were of primary importance. and its construction had been precisely controlled. The malerials used in the construction are locally available. The characteristics of the malerials. the construclion procedures. and the stability analysis are presented in detail. Ii) 1999 IAWQ Published by Elsevier Science Ltd. All rights reserved.
KEYWORDS Embankment; marl; recharge; sand; stability; wastewater. INTRODUCTION The Kingdom of Saudi Arabia has witnessed massive and unparalleled urbanization and agncultural development in the last 2-3 decades. This is reflected in the increase in the demand for water. In the sixth development plan (Sixth Development Plan, 1994), the Ministry of Planning put the water demand in 1994 for municipal and industrial purposes at 1800 million cubic metres (MCM). This was projected to grow to 2800 MCM by the year 1999. Most of this water returns as wastewater to the kingdom's various wastewater treatment plants. Such effluents are treated to various levels and then discharged into the Red Sea in western Saudi Arabia and the Arabian Gulf in eastern Saudi Arabia or into the desert. Of the estimated 1000 MCM of wastewater generated in the kingdom, only about 150 MCM is reused. Furthermore, the volume of wastewater generated is expected to grow to 1500 MCMlyr by the year 2000. 25
26
S. A. AIBAN et at.
This uncontrolled discharge of wastewater in the aforementioned methods has many environmental consequences and could easily lead to degradation of groundwater quality. A concern associated with the disposal oftreated or untreated sewage on or below the land surface revolves around the question of how far and how fast pathogenic bacteria and viruses can move in subsurface flow systems (Freeze and Cherry, 1979). Wellings et al. (1975) demonstrated vertical and lateral movement of viruses in secondary effluent discharged into a cypress dome. Viruses were shown to migrate 7-38 m laterally from the application point and to survive at least a period of 28 days (Ishaq et al., 1997). It is therefore necessary to have an environmentally friendly approach to dispose of treated wastewaters generated in the kingdom. One promising technique is the controlled recharge of aquifers using treated wastewaters. The process of recharge also results in the removal of contaminants from the wastewaters. In addition, the recharged water could be used to supplement the current water resources of the kingdom as it can be used for irrigation. Thus, wastewater recharge would not only provide an environmentally friendly means for disposal of treated wastewater, but would also contribute to the increased reuse of wastewater.
Wastewater recharge offers several advantages. It is more economical to recharge and use the recharged water than to use water from other sources such as desalination or other traditional tertiary treatment techniques. Furthermore, the aquifers into which the wastewater is recharged can serve as multi-year reservoirs. Underground reservoirs are ideal storage facilities in desert-type climates where evaporation rates are high. The use of reclaimed wastewater for agriculture may also lead to a reduction in the amount of commercial fertilizers applied. This is due to the presence of nitrogen and phosphorus compounds in the reclaimed water (Moore et al., 1985). In addition, it is a secured source of supply even during times of drought (Guymon and Hromadka, 1985). METHODS AND RESULTS
Treated wastewater can be recharged either through injection wells or through spreading basins. Spreading basins are frequently used. This involves the surface spreading of water in spreading or recharge basins and has the added advantage of filtering the water during the recharge process. Recharge or injection wells, on the other hand, are used to directly recharge water into deep water-bearing zones and into confined aquifers. Injection/recharge wells are the only alternative (Zikmund and Cole, 1996) where land is scarce and large areas for spreading cannot be made available. Although there are numerous cases of recharge wells being used for secondary wastewater recharge, spreading basins present fewer problems related to clogging, and the maintenance is cheaper. Desert soils in the kingdom have high infiltration and percolation rates, making them quite suitable for spreading basins. Surface spreading is most effective where there are no impending layers between the land surface and the aquifer. A potential site within the Al-Khobar wastewater treatment plant in eastern Saudi Arabia has been selected. The necessary site investigation has been performed, including a 20 m deep borehole. The site was found to be adequate for the field trial. Due to the shallowness of the groundwater table, it was decided to construct the filter layer (dune) above the natural ground level. An embankment was constructed to contain the sand filter. The details of material selection, design, analysis, and construction are presented below. Material selection for the dune and the retaining embankment The sand filter consists of clean dune sand confined within an embankment made of compacted marl/sand mixes. A typical cross-section of the dune and the retaining embankment (marl/sand material) is shown in Fig. 1. The sand used in the construction of the filter was a typical eastern Saudi poorly graded fine to medium sand, which was obtained from a dune on the Dhahran-Abqaiq road about 40 km west of Dhahran. There were not many choices, with regard to the grain size or composition of the sand, because all dune sands in the area have somewhat similar characteristics (Aiban et al., 1995). The grain size distribution of the sand is shown in Fig. 2.
27
Utilization of local material in the construction of an embankment
Effluent
Geomembrane Concrete Walkway
New Dune
Figure 1. Typical cross-section of the dune and the retaining embankment (wells are not shown).
Sieve No.
....
...... 000
"''''M
r
100 80
r
OJ)
.Sen
en ell p..
~ Q) ~
j,... . .
60
...- )
40
Q)
l:l...
20
o
)
0.01
0.10
I- - •.
.-
-0-
-
/'
/
7
marl (
~ sand
1.00
10.00
100.00
Grain size (rom) Figure 2. Grain size distribution for the sand and marl.
The marl used for the retaining structure was obtained from the same area as the sand, adjacent to the Dhahran-Abqaiq road. The selection of the embankment material was considered critical for this project because permeability as well as strength are of equal importance. A blend of marl and sand should provide a retaining structure that will withstand imposed loads and will provide a minimum permeability to give a 1-0 vertical flow through the dune. In the selection process, the physical properties of the marl were determined. These include gradation, plasticity, and compaction characteristics. The grain size distribution of the marl is shown in Fig. 2. The liquid and plastic limits (LL and PL) of the material passing US sieve No. 40 are 98 and 63, respectively. Thus, the plasticity index (PI) of the marl becomes 35, which is desirable since soils with plasticity have lower permeability when compared to non-plastic soils. On the other hand, it is known that for plastic soils, the strength decreases as the plasticity increases and the material becomes more water-sensitive. It was, therefore, decided to blend the marl with different percentages of sand and evaluate the resulting mixes for the desired permeability, strength, and swelling potential.
28
S. A. AIBAN et al.
The strength was obtained from California Bearing Ratio (CBR) tests, which were performed for each sample prepared for the compaction test. It is worth mentioning that all samples prepared for moisture• density and CBR tests were reconstituted to the gradation curve of the intended mix to reduce the variability. Samples were compacted using the modified Proctor test, which was performed according to the ASTM 0 1557 Method D standard, with an oversize correction. All CBR tests were performed on samples after 4 days of soaking and were performed in accordance with the ASTM D 1883 standard. This was helpful in comparing the strength of different sand/marl ratios prepared at different moisture contents. Test results clearly indicate that the maximum CBR values decrease as the marl content increases (Fig. 3). Marl content (%) 90
80
70
60
50
40
30 80
-;- 600 ~
....:
70
'-'
.<::
bh ..,c::
t;
.., .~ ..,'"
."" E 0
500
. ,,
400
I
300
U
..,
"0
c:: t.::: c::
200
,,
•
0 u
c:: :J
60
100 10
20
,-- .. -----
,
I
I
50
,, I,
40
40
U
30 20
CBR
30
'~ """'
l:t t:Q
10 50
60
70
Sand content (%) Figure 3. Variation of the strength with the sand/marl content for the blended material.
Due to the importance of volume change upon wetting, the amount of swelling, during the soaking period, was measured for all the samples prepared for the CBR tests. Test results show that swelling decreases as the molding moisture content increases. Furthermore, as the sand content increases the swelling decreases. Since the material is to be compacted on the wet side of optimum (to achieve low permeability), the swelling potential was found to be very low and will not be of concern, especially when using high sand content. In addition to the soaked CBR tests, unconfined compressive strength tests were performed on samples having a diameter of 100 mm and a height of 200 mm. All samples were reconstituted to their original gradation to reduce variability. Samples were compacted at a moisture content of 1-2% above their respective optimum moisture content values to insure compaction at a moisture content on the wet side of optimum, but close to the optimum values. This is known to give high strength at minimum permeability. The variations of the unconfined compressive strength (DCS) with sand content of the blend are shown in Fig. 3. It is clearly seen that a blend having a sand content of 30-40% gives the highest UCS values, which are slightly more than 500 kPa. Such blends need to be evaluated and compared, taking into account the strength and corresponding penneability values.
Since the primary objectives of the embankment are to retain the sand (filter) and minimize the lateral flow of effiuents through the sides of the embankment, the penneabilities of the different blends were evaluated. Careful inspection of the results shown in Fig. 4 will lead to the conclusion that a blend having about 55% sand and 45% marl gives reasonably high strength at relatively low permeability. This ratio was adopted for field construction of the embankment. It is worth mentioning that all samples for penneability testing were compacted in the CBR molds at a moisture content of 1-2% higher than the corresponding optimum moisture content values. All samples for permeability testing have a thickness of 46.4 mm, and the test was performed under a constant head of3850 mm. It is clearly seen that the permeability values are low for all tested sand/marl ratios. However, it is noticed that the penneability increases as the sand content increases,
Ctilization of local material in the construction of an embankment
29
but the maximum value (6* 1O~1O m/sec) is very close to the values required for clay liners. It is expected that the field values will be somewhat higher due to inclusion of oversize particles (larger than 19 mm aggregate), and field compaction may not be the same as the laboratory "controlled" compaction. Furthermore, the homogeneity of the material may induce some variation in the propertie~.
-+[_ . ._. _
7.0E-1O
Permeability
'.
lJncoofined ~ive
_
_~
'.
,i--,
_
~
J
---,- -- - 600
I
I
6.0E-1O -- -- - -(--\- - -"',- I
5.0&10
- -
500
............
,,
400
4.0&10
\!
3.0&10
.
~ ~\
I
\
1-
2.0E-1O 1.0E-1O
300
,
\
\
\
•
200 100
10
20
30
40
50
60
70
Sand content ('Yo) Figure 4. Variation of the permeability and unconfined compressive strength with sand content for the blended material.
Slope stability analysis The strength of the retaining embankment is required so that slope failures are prevented under the worst loading conditions. The slope stability analysis was performed using a computer program based on the generalized limit equilibrium (GLE) method. The program used was the 1995 version of Slope/W, which was developed by Geo-Slope International. The analysis was made assuming two dimensional analysis and rotational failure with circular sliding surfaces. Different scenarios were tried using different geometries with undrained shear strength parameters. The parameters considered include side slopes and height of the embankment, inclusion of reinforcing geosynthetic layers and their position. berm width, and height and width of the embankment at the top. The shear strength parameters assumed during the analysis are based on the Mohr-Coulomb failure criteria with a cohesion of c = 40 kPa and an angle of internal friction of, = 15°. To account for the worst field situation and soaking effect, these strength parameters are lower than those obtained from the laboratory triaxial testing. Different limits for the grid of the center of the slip surface and different tangents to the slip surface were tried. The factor of safety against sliding was computed for 1210 trial slip surfaces for each case. A factor of safety of 1.2 was assumed appropriate for this situation and will be achieved when a berm around the perimeter of the embankment is provided. The berm will extend for about 3 meters and have a height of about 2 meters above the natural ground level. The slip surface with the minimum factor of safety based on the GLE method is shown in Fig. 5. An acceptable factor of safety could also be obtained jf layers of geotextilcs or geogrids are included; however, the inclusion of such layers will affect the flow of effluent and complicate the numerical seepage analysis.
30
s.
A. AlBA
cl
"I.
1206· r--
Horizontal Distance (m) Figure 5. Failure surface (hased on 30 slices) and the rcsultlllg minimum lactor of safety
t(n
the emhankment.
Constmction of the sand dune and supporting embankment An area of 39 x 39 m was demarcated initially in the field and was excavated to the level of the water table, which is approximately 1.6 III below ground surface (GS). The positions of the monitoring wells were established, and they were erected. The entire excavated area was then filled with dunc sand to GS. Thereafter, the embankment and the dunc were constructed simultaneously. Thc dunc sand for the filter was dumped wilhout compaction to cover an area of 25 x 25 Ill. However, the embankment was compacted and the compaction was done on lins of about 30 Clll each. The material used for the emhankment is a blend of 45% marl and 55% sand. The degree of compaction was checked for each layer using a nuclear moisture density gauge. All layers were compacted to at least 95% of maximum dry dcnsity, according to the modified Proctor rcsults, and at moisture content on the wet side but close to optimum. In addition to the compaction tests, samples for the evaluation of the field penllcability of the embankment were taken. The samples were ohtained by pushing thc CBR mold into the compaetl.-d layer. Samples were then subjected to a constant head penneahility test. The average permeability for all tested spceimens (from three different layers) was found to he 3.4*10 K m/s. This value is much higher than the laboratory prepared samples. This could be altrihuted to the following: I. The laboratory samples were prepared IIndcr controlled comlilions and. thlls. can be assumed homogenous and properly compacted. 2. The field samples were compacted using a smooth drum compactor. which usually rcsults in higher pemleahility when compared to dynamic lahoratory compaction. 3. Field density and compaction moisture content are somewhat variahle and a highcr degrce of compaction cannot easily be achieved. However, the pcmleability is expected to decrease with compaction of the subscqucnt laycrs due to the compaelion efl()rts and the overhurden provided by the new soil laycrs. Furthennore. the penneahilily valucs afe expected 10 decrease due to clogging by the effluent. 11 is worth mcntioning that the pemleahilily
Utilization of local malerial in the construclIon of an embankment
31
of the sand filter is 1.4* I0- 2 mls at the placement void ratio. This value is much higher than that of the embankment. Performance of the embankment Upon completion of the embankment construction and the erection of all monitoring systems, a recharge run was initiated. The effluent was provided through a network of pipes on thc surface of the filter (dune) as shown in Fig. 6. The pumping system was fully automated whereby the discharge into the dune continued until the water head was 0.3 m above the dune's surfacc. The first trial took 3 days for the water to appear on the ground surface. During this period. things looked normal. Howevt:r, the discharge was not stopped when the water accumulated around the beml. a situation that should have been avoided. Such accumulation reduced the factor of safety of the embankment in general and that of the benn in particular. Upon the continuation of the discharge. the groundwater table rosc approximately 300 mm above the natural ground surface and the seepage from the monitoring wells eroded the berm. Local slides and slope failures in the berm were observed in a few places. However. the embankment was not affected except for minor cracks at the top. These slope failures and slides were triggered by the piping and excessive seepage pressure initiated at the bOllom of the monitoring wells. This was enhanced by the low dcgree of compaction around the wells and towards the edge of the berm due to lack of confinement. It is worth mentioning that a total of 29 wells have been constructed: eight of which are within the edge of the berm. The bases of the wells are placed within coarse aggregate filter to insure a free draining system. The high pore pressure induced high seepage forces at the base of the wells, which was subsequently followed by piping and washout of the adjacent berm material. The discharge was stopped and the damage in the benn was fixed. Subsequent runs did not result in any damage. but the groundwater table was not allowed to rise above the natural ground surface. This is what had been assumed in the design and slope stability analysis. Furthermore. no lateral seepage was observed in the embankment. It is expected that the present design will bt: adequate for the intended objectives.
Figure 6. Network of effluent distribution pipes on thc surlace Oflhc dunc.
CONCLUSIONS
Utilization of the available soils in eastern Saudi Arabia for the construction of the recharge facilities was f()und to serve the purpose adequately. FurthernlOre. the cost of Ihe embankmenl was much less than Ihat of eoncrctc retaining structures. A major advantage of the use of soils over contTele or sted structures is the elimination of corrosion of concrete and steel structures. especially in an environment where chloride and sulphale arc abundant. In atldition. there are no limilalions on size and. Iherelore. Ihere is no need for themlal or expansion joints. With regard to material selection. there is a need for a criteria whereby limits on strength and hytlraulic conductivity arc specifictl.
32
S. A. AlBAN et 01.
ACKNOWLEDGEMENTS The authors wish to acknowledge the support extended by King Abdulaziz City for Science and Technology (KACSn under Project No. AR-15-68 and the King Fahd University of Petroleum & Minerals for their continuous support. REFERENCES Aiban, S. A., AI-Abdul Wahhab, H. I. and AI-Amoudi, O. S. B. (1995). IdentificatIon, Evaluation and Improvement ofEastern Saudi Soils/or Constructional Purposes, Progress Report No.2, KACST AR-14-6I, submitted to King Abdulaziz City for Science and Technology, Riyadh, Saudi Atabia. Freeze, R. A. and Cherry, J. A. (1979). Groundwater. Prentice Hall Inc., New Jersey. Geo-Slope International Ltd (1995). User Guide for Siope/W. Calgary, Alberta, Canada. Guymon, G. L. and Hromadka, T. V. (1985). Modeling of groundwater response to artificial recharge. In: Artifictal Recharge of Groundwater, Takashi Asano (ed.), Butterworth Publishers. Isbaq, A. M., Khan, A. A., AI-Suwalyan, M. S., Farooq, S. and Alban, S. A. (1997). Contaminant removal using wastewater recbarge. Proc., Symposium on Civil Engineering and The Environment, Department of CIVil Engineering, King Fahd University of Petroleum & Mmerals, Dhahran, pp. 227-238. Moore, C. V., Olson, K. D. and Marino, M. A. (1985). On-fonn economics of reclaimed wastewater imgation. In: Irrigation with Reclaimed Mumclpal Wastewater - A GUidance Manual, Stuart Pettygrove and Takashi Asano (eds.), Lewis Publishers, Chelsea, Mich. Sixth Development Plan /415-/420 (1994). Mmistry of Planning Press, Riyadh, Saudi Atabia. Wellings, F. M., Lewis, A. L., Mountain, C. W. and Pierce, L. V. (1975). Appl Microbiol., 29, p. 1751. Zilcmund, K. S. and Cole, E. (1996). Artificial recharge: a water management tool. Int. Groundwater Tech., 4(2).
s
Pergamon
War. Sci. Tech. Vol. 40, No.7, pp. 25-32.1999 C 1999IAWQ Published by Elsevier Science LId
Prinled in Oreal Brilain. All rights reserved
0273-1223J99 $20.00 + 0.00
PD: 50273-1223(99)00580-6
UTILIZATION OF LOCAL MATERIAL IN THE CONSTRUCTION OF AN EMBANKMENT FOR RECHARGING GROUNDWATER AQUIFER WITH TREATED WASTEWATER EFFLUENTS S. A. Aiban, A. M. Ishaq and M. S. AI-Suwaiyan Department of Civil Engineering, King Fahd University ofPetroleum &< Minerals. Dhahran 31261. Saudi Arabia
ABSTRACT Water is certainly the moSt precious and valuable resowce of the physical environment for all living creatures. It is now well established that in many agricultural regions in the Kingdom of Saudi Arabia. the water table has been depleted dramatically. If the current level of agricultural production is 10 be maintained. other sources of agricultural water must be found. There is at least a billion cubic metres of secondary municipal effluent wasted annually. Thus. utilization of these wastewater effluents to recharge groundwater aquifers and reuse them at least for agricultural purposes becomes a viable proposition. In the study reported herein, treated wastewater will be used 10 recharge aquifers using spreading basins. The basin consists of a dune sand filter confined. from all sides. by an embankment of compacted sand/marl material. A field site has been constructed in AJ-Aziziyah in eastern Saudi Arabia. The sand filter is constructed from clean dune sand. has an area of 25 x 25 rn, and has an effective height of approximately 7.0 m. The stability and permeability of the embankment were of primary importance. and its construction had been precisely controlled. The malerials used in the construction are locally available. The characteristics of the malerials. the construclion procedures. and the stability analysis are presented in detail. Ii) 1999 IAWQ Published by Elsevier Science Ltd. All rights reserved.
KEYWORDS Embankment; marl; recharge; sand; stability; wastewater. INTRODUCTION The Kingdom of Saudi Arabia has witnessed massive and unparalleled urbanization and agncultural development in the last 2-3 decades. This is reflected in the increase in the demand for water. In the sixth development plan (Sixth Development Plan, 1994), the Ministry of Planning put the water demand in 1994 for municipal and industrial purposes at 1800 million cubic metres (MCM). This was projected to grow to 2800 MCM by the year 1999. Most of this water returns as wastewater to the kingdom's various wastewater treatment plants. Such effluents are treated to various levels and then discharged into the Red Sea in western Saudi Arabia and the Arabian Gulf in eastern Saudi Arabia or into the desert. Of the estimated 1000 MCM of wastewater generated in the kingdom, only about 150 MCM is reused. Furthermore, the volume of wastewater generated is expected to grow to 1500 MCMlyr by the year 2000. 25
26
S. A. AIBAN et at.
This uncontrolled discharge of wastewater in the aforementioned methods has many environmental consequences and could easily lead to degradation of groundwater quality. A concern associated with the disposal oftreated or untreated sewage on or below the land surface revolves around the question of how far and how fast pathogenic bacteria and viruses can move in subsurface flow systems (Freeze and Cherry, 1979). Wellings et al. (1975) demonstrated vertical and lateral movement of viruses in secondary effluent discharged into a cypress dome. Viruses were shown to migrate 7-38 m laterally from the application point and to survive at least a period of 28 days (Ishaq et al., 1997). It is therefore necessary to have an environmentally friendly approach to dispose of treated wastewaters generated in the kingdom. One promising technique is the controlled recharge of aquifers using treated wastewaters. The process of recharge also results in the removal of contaminants from the wastewaters. In addition, the recharged water could be used to supplement the current water resources of the kingdom as it can be used for irrigation. Thus, wastewater recharge would not only provide an environmentally friendly means for disposal of treated wastewater, but would also contribute to the increased reuse of wastewater.
Wastewater recharge offers several advantages. It is more economical to recharge and use the recharged water than to use water from other sources such as desalination or other traditional tertiary treatment techniques. Furthermore, the aquifers into which the wastewater is recharged can serve as multi-year reservoirs. Underground reservoirs are ideal storage facilities in desert-type climates where evaporation rates are high. The use of reclaimed wastewater for agriculture may also lead to a reduction in the amount of commercial fertilizers applied. This is due to the presence of nitrogen and phosphorus compounds in the reclaimed water (Moore et al., 1985). In addition, it is a secured source of supply even during times of drought (Guymon and Hromadka, 1985). METHODS AND RESULTS
Treated wastewater can be recharged either through injection wells or through spreading basins. Spreading basins are frequently used. This involves the surface spreading of water in spreading or recharge basins and has the added advantage of filtering the water during the recharge process. Recharge or injection wells, on the other hand, are used to directly recharge water into deep water-bearing zones and into confined aquifers. Injection/recharge wells are the only alternative (Zikmund and Cole, 1996) where land is scarce and large areas for spreading cannot be made available. Although there are numerous cases of recharge wells being used for secondary wastewater recharge, spreading basins present fewer problems related to clogging, and the maintenance is cheaper. Desert soils in the kingdom have high infiltration and percolation rates, making them quite suitable for spreading basins. Surface spreading is most effective where there are no impending layers between the land surface and the aquifer. A potential site within the Al-Khobar wastewater treatment plant in eastern Saudi Arabia has been selected. The necessary site investigation has been performed, including a 20 m deep borehole. The site was found to be adequate for the field trial. Due to the shallowness of the groundwater table, it was decided to construct the filter layer (dune) above the natural ground level. An embankment was constructed to contain the sand filter. The details of material selection, design, analysis, and construction are presented below. Material selection for the dune and the retaining embankment The sand filter consists of clean dune sand confined within an embankment made of compacted marl/sand mixes. A typical cross-section of the dune and the retaining embankment (marl/sand material) is shown in Fig. 1. The sand used in the construction of the filter was a typical eastern Saudi poorly graded fine to medium sand, which was obtained from a dune on the Dhahran-Abqaiq road about 40 km west of Dhahran. There were not many choices, with regard to the grain size or composition of the sand, because all dune sands in the area have somewhat similar characteristics (Aiban et al., 1995). The grain size distribution of the sand is shown in Fig. 2.
27
Utilization of local material in the construction of an embankment
Effluent
Geomembrane Concrete Walkway
New Dune
Figure 1. Typical cross-section of the dune and the retaining embankment (wells are not shown).
Sieve No.
....
...... 000
"''''M
r
100 80
r
OJ)
.Sen
en ell p..
~ Q) ~
j,... . .
60
...- )
40
Q)
l:l...
20
o
)
0.01
0.10
I- - •.
.-
-0-
-
/'
/
7
marl (
~ sand
1.00
10.00
100.00
Grain size (rom) Figure 2. Grain size distribution for the sand and marl.
The marl used for the retaining structure was obtained from the same area as the sand, adjacent to the Dhahran-Abqaiq road. The selection of the embankment material was considered critical for this project because permeability as well as strength are of equal importance. A blend of marl and sand should provide a retaining structure that will withstand imposed loads and will provide a minimum permeability to give a 1-0 vertical flow through the dune. In the selection process, the physical properties of the marl were determined. These include gradation, plasticity, and compaction characteristics. The grain size distribution of the marl is shown in Fig. 2. The liquid and plastic limits (LL and PL) of the material passing US sieve No. 40 are 98 and 63, respectively. Thus, the plasticity index (PI) of the marl becomes 35, which is desirable since soils with plasticity have lower permeability when compared to non-plastic soils. On the other hand, it is known that for plastic soils, the strength decreases as the plasticity increases and the material becomes more water-sensitive. It was, therefore, decided to blend the marl with different percentages of sand and evaluate the resulting mixes for the desired permeability, strength, and swelling potential.
28
S. A. AIBAN et al.
The strength was obtained from California Bearing Ratio (CBR) tests, which were performed for each sample prepared for the compaction test. It is worth mentioning that all samples prepared for moisture• density and CBR tests were reconstituted to the gradation curve of the intended mix to reduce the variability. Samples were compacted using the modified Proctor test, which was performed according to the ASTM 0 1557 Method D standard, with an oversize correction. All CBR tests were performed on samples after 4 days of soaking and were performed in accordance with the ASTM D 1883 standard. This was helpful in comparing the strength of different sand/marl ratios prepared at different moisture contents. Test results clearly indicate that the maximum CBR values decrease as the marl content increases (Fig. 3). Marl content (%) 90
80
70
60
50
40
30 80
-;- 600 ~
....:
70
'-'
.<::
bh ..,c::
t;
.., .~ ..,'"
."" E 0
500
. ,,
400
I
300
U
..,
"0
c:: t.::: c::
200
,,
•
0 u
c:: :J
60
100 10
20
,-- .. -----
,
I
I
50
,, I,
40
40
U
30 20
CBR
30
'~ """'
l:t t:Q
10 50
60
70
Sand content (%) Figure 3. Variation of the strength with the sand/marl content for the blended material.
Due to the importance of volume change upon wetting, the amount of swelling, during the soaking period, was measured for all the samples prepared for the CBR tests. Test results show that swelling decreases as the molding moisture content increases. Furthermore, as the sand content increases the swelling decreases. Since the material is to be compacted on the wet side of optimum (to achieve low permeability), the swelling potential was found to be very low and will not be of concern, especially when using high sand content. In addition to the soaked CBR tests, unconfined compressive strength tests were performed on samples having a diameter of 100 mm and a height of 200 mm. All samples were reconstituted to their original gradation to reduce variability. Samples were compacted at a moisture content of 1-2% above their respective optimum moisture content values to insure compaction at a moisture content on the wet side of optimum, but close to the optimum values. This is known to give high strength at minimum permeability. The variations of the unconfined compressive strength (DCS) with sand content of the blend are shown in Fig. 3. It is clearly seen that a blend having a sand content of 30-40% gives the highest UCS values, which are slightly more than 500 kPa. Such blends need to be evaluated and compared, taking into account the strength and corresponding penneability values.
Since the primary objectives of the embankment are to retain the sand (filter) and minimize the lateral flow of effiuents through the sides of the embankment, the penneabilities of the different blends were evaluated. Careful inspection of the results shown in Fig. 4 will lead to the conclusion that a blend having about 55% sand and 45% marl gives reasonably high strength at relatively low permeability. This ratio was adopted for field construction of the embankment. It is worth mentioning that all samples for penneability testing were compacted in the CBR molds at a moisture content of 1-2% higher than the corresponding optimum moisture content values. All samples for permeability testing have a thickness of 46.4 mm, and the test was performed under a constant head of3850 mm. It is clearly seen that the permeability values are low for all tested sand/marl ratios. However, it is noticed that the penneability increases as the sand content increases,
Ctilization of local material in the construction of an embankment
29
but the maximum value (6* 1O~1O m/sec) is very close to the values required for clay liners. It is expected that the field values will be somewhat higher due to inclusion of oversize particles (larger than 19 mm aggregate), and field compaction may not be the same as the laboratory "controlled" compaction. Furthermore, the homogeneity of the material may induce some variation in the propertie~.
-+[_ . ._. _
7.0E-1O
Permeability
'.
lJncoofined ~ive
_
_~
'.
,i--,
_
~
J
---,- -- - 600
I
I
6.0E-1O -- -- - -(--\- - -"',- I
5.0&10
- -
500
............
,,
400
4.0&10
\!
3.0&10
.
~ ~\
I
\
1-
2.0E-1O 1.0E-1O
300
,
\
\
\
•
200 100
10
20
30
40
50
60
70
Sand content ('Yo) Figure 4. Variation of the permeability and unconfined compressive strength with sand content for the blended material.
Slope stability analysis The strength of the retaining embankment is required so that slope failures are prevented under the worst loading conditions. The slope stability analysis was performed using a computer program based on the generalized limit equilibrium (GLE) method. The program used was the 1995 version of Slope/W, which was developed by Geo-Slope International. The analysis was made assuming two dimensional analysis and rotational failure with circular sliding surfaces. Different scenarios were tried using different geometries with undrained shear strength parameters. The parameters considered include side slopes and height of the embankment, inclusion of reinforcing geosynthetic layers and their position. berm width, and height and width of the embankment at the top. The shear strength parameters assumed during the analysis are based on the Mohr-Coulomb failure criteria with a cohesion of c = 40 kPa and an angle of internal friction of, = 15°. To account for the worst field situation and soaking effect, these strength parameters are lower than those obtained from the laboratory triaxial testing. Different limits for the grid of the center of the slip surface and different tangents to the slip surface were tried. The factor of safety against sliding was computed for 1210 trial slip surfaces for each case. A factor of safety of 1.2 was assumed appropriate for this situation and will be achieved when a berm around the perimeter of the embankment is provided. The berm will extend for about 3 meters and have a height of about 2 meters above the natural ground level. The slip surface with the minimum factor of safety based on the GLE method is shown in Fig. 5. An acceptable factor of safety could also be obtained jf layers of geotextilcs or geogrids are included; however, the inclusion of such layers will affect the flow of effluent and complicate the numerical seepage analysis.
30
s.
A. AlBA
cl
"I.
1206· r--
Horizontal Distance (m) Figure 5. Failure surface (hased on 30 slices) and the rcsultlllg minimum lactor of safety
t(n
the emhankment.
Constmction of the sand dune and supporting embankment An area of 39 x 39 m was demarcated initially in the field and was excavated to the level of the water table, which is approximately 1.6 III below ground surface (GS). The positions of the monitoring wells were established, and they were erected. The entire excavated area was then filled with dunc sand to GS. Thereafter, the embankment and the dunc were constructed simultaneously. Thc dunc sand for the filter was dumped wilhout compaction to cover an area of 25 x 25 Ill. However, the embankment was compacted and the compaction was done on lins of about 30 Clll each. The material used for the emhankment is a blend of 45% marl and 55% sand. The degree of compaction was checked for each layer using a nuclear moisture density gauge. All layers were compacted to at least 95% of maximum dry dcnsity, according to the modified Proctor rcsults, and at moisture content on the wet side but close to optimum. In addition to the compaction tests, samples for the evaluation of the field penllcability of the embankment were taken. The samples were ohtained by pushing thc CBR mold into the compaetl.-d layer. Samples were then subjected to a constant head penneahility test. The average permeability for all tested spceimens (from three different layers) was found to he 3.4*10 K m/s. This value is much higher than the laboratory prepared samples. This could be altrihuted to the following: I. The laboratory samples were prepared IIndcr controlled comlilions and. thlls. can be assumed homogenous and properly compacted. 2. The field samples were compacted using a smooth drum compactor. which usually rcsults in higher pemleahility when compared to dynamic lahoratory compaction. 3. Field density and compaction moisture content are somewhat variahle and a highcr degrce of compaction cannot easily be achieved. However, the pcmleability is expected to decrease with compaction of the subscqucnt laycrs due to the compaelion efl()rts and the overhurden provided by the new soil laycrs. Furthennore. the penneahilily valucs afe expected 10 decrease due to clogging by the effluent. 11 is worth mcntioning that the pemleahilily
Utilization of local malerial in the construclIon of an embankment
31
of the sand filter is 1.4* I0- 2 mls at the placement void ratio. This value is much higher than that of the embankment. Performance of the embankment Upon completion of the embankment construction and the erection of all monitoring systems, a recharge run was initiated. The effluent was provided through a network of pipes on thc surface of the filter (dune) as shown in Fig. 6. The pumping system was fully automated whereby the discharge into the dune continued until the water head was 0.3 m above the dune's surfacc. The first trial took 3 days for the water to appear on the ground surface. During this period. things looked normal. Howevt:r, the discharge was not stopped when the water accumulated around the beml. a situation that should have been avoided. Such accumulation reduced the factor of safety of the embankment in general and that of the benn in particular. Upon the continuation of the discharge. the groundwater table rosc approximately 300 mm above the natural ground surface and the seepage from the monitoring wells eroded the berm. Local slides and slope failures in the berm were observed in a few places. However. the embankment was not affected except for minor cracks at the top. These slope failures and slides were triggered by the piping and excessive seepage pressure initiated at the bOllom of the monitoring wells. This was enhanced by the low dcgree of compaction around the wells and towards the edge of the berm due to lack of confinement. It is worth mentioning that a total of 29 wells have been constructed: eight of which are within the edge of the berm. The bases of the wells are placed within coarse aggregate filter to insure a free draining system. The high pore pressure induced high seepage forces at the base of the wells, which was subsequently followed by piping and washout of the adjacent berm material. The discharge was stopped and the damage in the benn was fixed. Subsequent runs did not result in any damage. but the groundwater table was not allowed to rise above the natural ground surface. This is what had been assumed in the design and slope stability analysis. Furthermore. no lateral seepage was observed in the embankment. It is expected that the present design will bt: adequate for the intended objectives.
Figure 6. Network of effluent distribution pipes on thc surlace Oflhc dunc.
CONCLUSIONS
Utilization of the available soils in eastern Saudi Arabia for the construction of the recharge facilities was f()und to serve the purpose adequately. FurthernlOre. the cost of Ihe embankmenl was much less than Ihat of eoncrctc retaining structures. A major advantage of the use of soils over contTele or sted structures is the elimination of corrosion of concrete and steel structures. especially in an environment where chloride and sulphale arc abundant. In atldition. there are no limilalions on size and. Iherelore. Ihere is no need for themlal or expansion joints. With regard to material selection. there is a need for a criteria whereby limits on strength and hytlraulic conductivity arc specifictl.
32
S. A. AlBAN et 01.
ACKNOWLEDGEMENTS The authors wish to acknowledge the support extended by King Abdulaziz City for Science and Technology (KACSn under Project No. AR-15-68 and the King Fahd University of Petroleum & Minerals for their continuous support. REFERENCES Aiban, S. A., AI-Abdul Wahhab, H. I. and AI-Amoudi, O. S. B. (1995). IdentificatIon, Evaluation and Improvement ofEastern Saudi Soils/or Constructional Purposes, Progress Report No.2, KACST AR-14-6I, submitted to King Abdulaziz City for Science and Technology, Riyadh, Saudi Atabia. Freeze, R. A. and Cherry, J. A. (1979). Groundwater. Prentice Hall Inc., New Jersey. Geo-Slope International Ltd (1995). User Guide for Siope/W. Calgary, Alberta, Canada. Guymon, G. L. and Hromadka, T. V. (1985). Modeling of groundwater response to artificial recharge. In: Artifictal Recharge of Groundwater, Takashi Asano (ed.), Butterworth Publishers. Isbaq, A. M., Khan, A. A., AI-Suwalyan, M. S., Farooq, S. and Alban, S. A. (1997). Contaminant removal using wastewater recbarge. Proc., Symposium on Civil Engineering and The Environment, Department of CIVil Engineering, King Fahd University of Petroleum & Mmerals, Dhahran, pp. 227-238. Moore, C. V., Olson, K. D. and Marino, M. A. (1985). On-fonn economics of reclaimed wastewater imgation. In: Irrigation with Reclaimed Mumclpal Wastewater - A GUidance Manual, Stuart Pettygrove and Takashi Asano (eds.), Lewis Publishers, Chelsea, Mich. Sixth Development Plan /415-/420 (1994). Mmistry of Planning Press, Riyadh, Saudi Atabia. Wellings, F. M., Lewis, A. L., Mountain, C. W. and Pierce, L. V. (1975). Appl Microbiol., 29, p. 1751. Zilcmund, K. S. and Cole, E. (1996). Artificial recharge: a water management tool. Int. Groundwater Tech., 4(2).
s