IJOER-MAY-2017-24

Views:
 
     
 

Presentation Description

Optimal Well Locations using Genetic Algorithm for Tushki Project, Western Desert, Egypt

Comments

Presentation Transcript

slide 1:

International Journal of Engineering Research Science IJOER ISSN: 2395-6992 Vol-3 Issue-7 July- 2017 Page | 1 Optimal Well Locations using Genetic Algorithm for Tushki Project Western Desert Egypt S. Khalaf 1 M.I. Gad 2 1 Irrigation and Hydraulics Dept. Faculty of Engineering Mansoura University Egypt. 2 Prof. Dr. Hydrology Division Desert Research Center Cairo Egypt Abstract — The well location problem is challenging due to the non-linear discrete and often multi-modal objective function. The optimal distribution of productive well locations mitigates the groundwater productivity problem that threats the national projects in arid countries like Egypt. In this paper a trial to mitigate this problem in Tushki National Project south western desert Egypt is supposed via delineating the optimal well locations and optimal pumping rates. The methodology of combination between simulation and optimization techniques is applied. The simulation outputs of groundwater flow system by Visual MODFLOW model is linked by the constructed Fortran Code for Optimal well Location using Genetic Algorithm OLGA Code for obtaining the optimum management of groundwater resources in this project. Two management cases are considered by running the model domain with adopted steady and transit calibrated parameters. The first management case deals with the present well locations and predicts the optimal value of the objective function maximum pumping rate. In the second case the optimal new well locations resulted from the OLGA Code is predicted from flexible well location with the moving well option. Also the prediction of the future changes in both head and flow are studied in steady and transient states. Keywords — Groundwater Optimization Genetic Algorithm Tushki project. I. INTRODUCTION Selecting optimal well sites is a valuable problem to solve maximizing groundwater recovery increases oil reserves and minimizing costs improves profitability. One of the best methods of selecting optimal sites is the combined use of simulation and optimization S/O models. While simulation models basically provide solutions of the governing equations of groundwater flow optimization models identify an optimal management and planning strategy from a set of feasible alternative strategies. The genetic algorithm GA is widely used to modify the parameters of groundwater flow models Yan et al. 2003 Yao et al. 2003 and to solve the management models of groundwater resources Mckinney and Lin 1994 Liu et al. 2002 Zhu et al. 2003. Rana et al 2008 conducted a study on a spatio-temporal optimization of agricultural drainage using groundwater model and GA. Guan et al. 2008 proposed an improved genetic algorithm IGA to solve optimization problems with equality and inequality constraints. Hamid et al. 2009 presented a paper focusing on the S/O for conjunctive use of surface water and groundwater on a basin-wide scale the Najafabad plain in west-central Iran. Gad et al. 2011 and Moharram et al. 2012 used an optimization model based on the combination of MODFLOW simulation with GA to maximize the total pumping rate from the Nubian sandstone aquifer in El-Dakhla and El-Farafra depressions Egypt. Saafan et al. 2011 applied multi-objective genetic algorithm MOGA Code in El-Farafra depression Egypt to study the maximum pumping rate and minimum operation cost. Gad and Khalaf 2013 used MOGA model to develop the maximum pumping rate and minimum operation cost in the Miocene aquifer of Wadi El-Farigh West Delta Egypt. In the other side there are limited studies for the determination of optimal operating strategy including unknown well locations and pumping rates for groundwater systems to the best of our knowledge Wang and Ahlfeld 1994 Karahan and Ayvaz 2005 and Ayvaz and Karahan 2008. Saffi and Cheddadi 2007 developed an algebraic expression which gave the matrix of transient influence coefficients for one-dimensional semi-confined aquifer model and solved the governing equation by using a mixed compartment model. Tung and Chou 2004 integrated pattern classification and tabu search to optimize the average zonal groundwater pumping for an aquifer. As a general the research conducted in this field dealt with the identification of locations and released histories of unknown groundwater pollution sources Mahar and Datta 2000 2001 Aral et al. 2001 Ruperti 2002 Singh et al. 2004 Sun et al. 2006 pumping well optimization for optimum remediation design Wang and Ahlfeld 1994 Huang and Mayer 1997 Guan and Aral 1999 Zheng and Wang 1999 Mantoglou and Kourakos 2007 Chang et al. 2007 and optimum well locations and pumping rates in the coastal sides Park and Aral 2004 Ferreira da Silva and Haie 2007. Meira et al. 2016 proposed technique for decision-making process in optimal well location including a step of risk analysis associated with the uncertainties present in the variables of the

slide 2:

International Journal of Engineering Research Science IJOER ISSN: 2395-6992 Vol-3 Issue-7 July- 2017 Page | 2 problem. The proposed technique was applied to two benchmark cases and the results evaluated by experts in the field indicated that the obtained solutions were richer than those identified by previously adopted manual approaches. Due to the few of scientific researches related to optimal well locations especially in our national governmental and invest mental agricultural projects a trial was carried out in this paper to through a light on the optimal location for new fourteen production wells in Tushki project required for irrigation of 25 thousand acres which will drill under the strategy of reclamation of 1.5 million acres project. II. DESCRIPTION OF CASE STUDY 2.1 Site description Tushki area is a part of the Western Desert of Egypt between latitudes 22° 14 24.75" and 22° 50 6.10" N and longitudes 31° 0 44.06"and 31° 50 4.05E with total area of 2268 km 2 . Tushki project one of the national governmental and invest mental agricultural projects aims to reclaim 540000 acres using both surface water from El-Sheikh Zaid artificial canal 1.5 km 3 /year and groundwater extracted from about 316 productive wells for cultivating 36000 acres. Moreover the agricultural development in the north and south of Abu-Simbel area which depends only on the groundwater resource will extract 22.5 million m 3 /year from the present 75 productive wells Figure 1. This operation policy is expected to have its side effect on groundwater regime of Nubia Sandstone Aquifer in Abu-Simbel Area NSAASA due to the absence of optimal well locations which will be the fulcrum of this research. 2.2 Geomorphological and geological settings Referring to the geomorphological studies prepared by El-Shazly et al. 1980 CONOCO 1987 and Aggour et al. 2012 five geomorphologic units are distinguished in the study area. Nasser lake the Nile valley Wadi Kurkur pediplain Tushki depression and West Dungul plain. Wadi Kurkur pediplain represents most parts of the area. The pediplain surface is covered mainly by nearly horizontal beds of Nubia sandstone with outcrops of igneous and metamorphic rocks as well as several volcanic exposures. Tushki depression area is an elongated featureless smooth plain bounded by the topographic contour line of 200m above sea level and is generally representing a topographic depression with an average absolute elevation of less than 200m. It is bounded from the north and east by limestone plateau and from the south and the west by sandy pediplain and peneplain respectively. FIGURE1. LOCATION MAP OF ABU SIMBEL AREA SHOWING THE LOCATION OF 75 PRODUCTIVE WELLS AND 9 OBSERVATION WELLS Geologically the main geological units in Tushki area are the Precambrian Basement rocks and sedimentary succession. The basement rocks constitute some of the conspicuous ridges and high peaks in the area such as G. Umm Shaghir 318 m asl. The lithostratigraphic sedimentary succession is dominated by Cretaceous Oligocene and Quaternary sediments Figure 2. The Lower Cretaceous rocks are differentiated into the Abu Simbel Lake Nasser and Sabaya formations. The Abu Simbel Formation covers vast parts south of Khour Tushki with thickness ranges from 30m to 87.45m Aggour et al. 2012. The Lake Nasser Formation is underlain by the Abu Simbel Formation and overlain by the Sabaya Formation. In addition the Upper Cretaceous rocks are represented by Kiseiba Formation with 150m thick. The Oligocene basalt rocks are widespread near the head of Khour Tushki. Moreover the Quaternary deposits are represented by sand accumulations and Alluvial sediments. As a general various structural elements are present among which are the uplift of the basement Nakhlai-Aswan uplift trending NE-SW with 300 km long E-W faults north of Khour Tushki and NW-SE faults south of Khour Tushki 0 2 4 km MEDITERRANEAN SEA RED SEA SINAI Cairo Assiut Qena Aswan Lake Nasser Siwa Qattara Depression El- Bahariya El- Farafra El- Dakhla El- Kharga W E S T E R N D E S E R T E A S T E R N D E S E R T Gilf Kebir Plateau 25 00 30 00 35 00 22 00 26 00 30 00 0 100 200 km Alexandria Scale Study area. East El Oweinat o o o o o o Observation well

slide 3:

International Journal of Engineering Research Science IJOER ISSN: 2395-6992 Vol-3 Issue-7 July- 2017 Page | 3 E-W and NE-SW folds the Ring forms and the structural lineation. The fault systems play an important role in recharging the NSAASA from Lake Nasser especially the NW-SE. They act as conduits and help in transmitting the groundwater within the NSAASA supporting the hydraulic connection between the aquifer units Aggour et al. 2012. 2.3 Hydrogeological aspects The NSAASA is mainly composed of three successive water-bearing units namely the Gilf Abu Simbel and Lake Nasser. The middle unit is an aquiclude that extends under the lake and the lower unit overlying the granitic basement is a confined aquifer of fluvial sandstones Figure 3. The groundwater of NSAASA exists under unconfined to semi-confined conditions. The NSAASA maximum thickness reaches 400 m and decreases towards north and NW. The transmissivity of the NSAASA ranges between 389 and 1322 m 2 /day Table 2 and increases from SW to NE. The constructed equipotential contour maps Figure 4 show that the local recharge of the NSAASA depends upon the leakage from the Lake Nasser due west direction. The estimated groundwater flow rate based on these contour maps Figure 4 reaches 0.054 m/day near Nasser Lake and decreases to 0.044 m/day towards northwestern parts. FIGURE 3: NSAASA POTENTIALITY MAP AND SW-NE CROSS-SECTION SHOWING AQUIFER THICKNESS FIGURE 2: GEOLOGICAL MAP OF THE STUDY AREA AFTER CONCO 1987

slide 4:

International Journal of Engineering Research Science IJOER ISSN: 2395-6992 Vol-3 Issue-7 July- 2017 Page | 4 III. MATERIALS AND METHODS 3.1 Field trip and data collection The materials used in this paper are collected through carrying out four field trips in Abu Simbel area during the period 2010-12. The basic hydraulic parameters of NSAASA are estimated based on carrying out one step test and 13 long duration pumping tests Tables 1 and 2. The different hydrologic data are obtained during the field trips. These materials also include collection of archival data well drilling reports WRRI 2002 registration of discharge distribution of wells proposed operating systems for both groundwater supply and reclaimed area beside recording depth to water for groundwater level changes. TABLE 1 RESULTS OF STEP DRAWDOWN TEST ANALYSIS IN NSAASA Step No Step time Duration hour Well discharge Q m 3 /day Total DD s m s/ Q day/m 2 I 1.5 1080 8.32 0.008666667 II 1.5 1200 11.8 0.009833333 III 1.5 1368 13.38 0.009291667 IV 1.5 1488 15.82 0.01063172 TABLE 2 RESULTS OF LONG DURATION PUMPING TESTS ANALYSIS IN NSAASA Well No Pumping rate m 3 /h Depth to water m Total DD m Screen m Hydraulic parameters K regional m/day K local m/day T m 2 /day W27 105 79.4 5.67 80.5 6.88 12.45 1003 W30 120 76.79 7.89 80.5 4.9 9.09 732 W38 173 60.88 11.81 80.5 4.94 10.07 812 W40 120 62.59 10.01 80.5 9.21 16.42 1322 W43 112 71.26 - 80.5 - - 878 W45 160 55.6 16.28 80.5 5.19 10.41 839 W50 172 30.64 17.53 74.75 2.19 5.64 422 W55 123 58.6 24.87 57.5 - 9.28 534 W 57 110 63.15 20.84 57.5 - 8.06 464 W 60 100 74.07 13.17 135 - 7.11 389 W75 107 72.7 14.5 57.5 - 10.01 576 W80 110 68.8 11.25 57.5 - - 912 W85 130 58.2 - 57.5 - - 340 31 30 31 30 31 45 31 45 22 30 22 30 23 00 23 00 22 45 22 45 0 5 10 km Lake Nasser m a.s.l 110 120 130 140 150 160 170 FIGURE 4: THE GROUNDWATER EQUIPOTENTIAL MAPS OF NSAASA DURING 2008 LEFT MAP AFTER EL-SABRI ET. AL. 2010 AND DURING 2010 RIGHT MAP-AFTER AGGOUR ET. AL. 2012

slide 5:

International Journal of Engineering Research Science IJOER ISSN: 2395-6992 Vol-3 Issue-7 July- 2017 Page | 5 XX Y Y ZZ s h h h h K K K W S x x y y Z Z t                                  The methodological approach used in this paper is based on the mathematical modeling techniques which combines between simulation and optimization model. A computer programming with FORTRAN language has been originally established to apply the principles of the genetic algorithm for studying the optimal location of the groundwater productive wells as a principal tool for groundwater resources management. The Optimal well Location applying Genetic Algorithm OLGA Code links MODFLOW with genetic algorithm technique to establish a simulation optimization groundwater model. 3.2 Simulation technique Groundwater flow simulation is carried out applying MODFLOW software. It describes groundwater flow of constant density under non-equilibrium conditions in three-dimensional heterogeneous and anisotropic medium according to the following equation Bear 1979: 1 in which K xx K yy and K zz are values of hydraulic conductivity L T -l along the x y and z coordinate axes h is the potentiometric head L W is the volumetric flux per unit volume and represents sources and/or sinks of water T -l S s is the specific storage of the porous material L -1 and t is time T. 3.2.1 Conceptual Model The conceptual model of the NSAASA is based on the geology of the study area which is comprised of Nubian sandstone sediments. The hydrogeologic system is concerned two hydraulically connected through fault planes layers. The top hydrostratigraphic unit Gilf Formation is considered to be a sandy unconfined aquifer zone up to 95 meters maximum thick. The second confined aquifer zone of fluvial sandstones is separated from the sandy unconfined aquifer zone by aquiclude and had maximum thickness of 79 meters. The scattered new communities around the asphaltic road in the N-S direction of the Abu Simbel area are primarily planned for agriculture. 3.2.2 Model domain The model domain of NSAASA is selected to cover 3315 km 2 51x65 Km-Figure 3. It is discretized using 130 rows × 102 columns rectangular cells. This discrimination produces 13260 cells in the model layer. The width of the cells along rows in x-direction is 500m and along columns in y-direction is 500 m. The grid geometry is shown in Figure 5. 3.2.3 Boundary conditions The hydraulic conditions at the boundaries of the model domain are specified Figure 5. Based on the groundwater flow pattern of the NSAASA Figure 4 and the field measurements of the groundwater head in the present productive wells Table 2 the initial and boundary conditions of the local model is defined. The lower boundary water body of Lake Nasser is considered a time-constant head at level of 175 m asl DRC 2012. The equipotential contour line of 130m asl characterizing to the upper boundary Figure 5 is far enough from the well fields to be genral head boundary condition. The left and right boundaries are assumed to be no flow boundaries due to the regional flow direction from SE to the NW. The impermeable Basement rock is assumed to be the bottom boundary of the model domain with no flow and maximum depth 236.45m Table 3 and Figure 6. The hydraulic heads and hydraulic conductance are assigned to boundary cells. These heads vary during the simulation process according to different stresses applied on the modeled area. 3.2.4 Aquifer properties The aquifer material properties of the NSAASA include the transmissivity the storage coefficient and the hydraulic conductivity. The transmissivity ranges from 389 and 1322 m 2 /day Table 2 and the storage coefficient from 1.2x10 -4 to 5x10 -4 El-Sabri et. Al. 2010. Vertical heterogeneity is assumed adequate to allow treatment as multi layered aquifer. The vertical hydraulic conductivity estimated from the aquifer thickness Figure 6-left and aquifer transmissivity values is varied from 5.64 to 16.42 m/day. Furthermore the availability of a relatively complete database of 75 groundwater drilled wells and 9 observation wells in the NSAASA help in constructing the present model. In addition the initial piezometric heads are computed from the absolute ground elevation values of the model domain extracted from the Digital Elevation Model DEM of the study area Figure 6-right.

slide 6:

International Journal of Engineering Research Science IJOER ISSN: 2395-6992 Vol-3 Issue-7 July- 2017 Page | 6 TABLE 3 UPPER SURFACE OF BASEMENT ROCK AMSL IN ABU SIMBEL AREA Well name Ground elevation m Basement level m Well name Ground elev. m Basement level m Akaria well 185.07 -34 O.B No11 85 -236.45 Well No 4 193.57 -176 O.B No 3 225 +107 Well No 5 204.72 -176 O.B No 3 225 0.0 Well No 13 194.05 +10 O.B No 6" 201 -129 Well No 21 195 -4 O.B No 7 165 -86 KADCO No 1 230 -2 O.B No 8 215 -9 KADCO No 2 233 +8 O.B No 10" 165 +75 KADCO No 3 233 +15 O.B No 18 190 -42 El Masria well 164 +92 O.B No 18" 190 -116 Akaria well 2 198.06 -124 O.B No 19 193 0.0 O.B No 4 193.05 -168 O.B No 20 165 +51 O.B No 10 186.07 -45 O.B No 25 183 -41 FIGURE 5. FLOW MODEL DOMAIN GRID AND THE BOUNDARY CONDITIONS OF THE NSAASA MODEL

slide 7:

International Journal of Engineering Research Science IJOER ISSN: 2395-6992 Vol-3 Issue-7 July- 2017 Page | 7 FIGURE 6: AQUIFER THICKNESS CONTOUR MAP LEFT MAP AND GROUND ELEVATION EXTRACTED FROM DEM RIGHT MAP 3.2.5 Model calibration Calibration involves modification of the parameters and forcing functions in the model in such a way as to assure that the model reproduced observed-head values Hill 1998 2000. The calibration process typically involves calibrating to steady- state and unsteady state conditions. The performed steady state calibration permits the adjustment of the hydraulic conductivity where NSAASA storage changes are not significant. The relation between the calculated and observed heads is checked from the calculated-observed head curve and the variance appears as 45.966 Figure 8. This large value indicates great difference between the heads calculated by the model and actual measured heads. After many times of changing the k values the variance between the observed and calculated heads is minimized to 3.61 and the calculated head values is very closely related to the field-measured head values Figure 7. Based on the calibrated model the estimated hydraulic conductivity values are ranging from 4.33 to 7.60 m/day. These values are generally more slightly than the values calculated from the field data. The difference may relate to the benchmarks uncertainty and grid refinement. FIGURE 7: THE CALCULATED AND OBSERVED HEADS OF THE MODEL DOMAIN OF NSAASA FOR STEADY STATE Transient state specific yield values are modified on a trial and errors basis until a good match between the observed heads in the years 2002 and 2012 sited in Kim and Sultan 2002 and Aggour et. Al. 2012 and the calculated heads are achieved. The range of the resulted specific yield after the final calibration of the transient state is found to be varying from 0.14 to 0.34 Fig. 8. It can be seen that in general there is a good agreement between the observed and simulated head values.

slide 8:

International Journal of Engineering Research Science IJOER ISSN: 2395-6992 Vol-3 Issue-7 July- 2017 Page | 8 1 w N jj j Z Q P    FIG. 8: THE CALCULATED AND OBSERVED HEADS OF THE MODEL DOMAIN OF NSAASA FOR UNSTEADY STATE 3.3 Optimization Technique Genetic Algorithms GA are adaptive heuristic search algorithm premised on the evolutionary ideas of natural selection and genetic. The basic concept of GA is designed to simulate processes in natural system necessary for evolution. A solution for a given problem is represented in the form of a string called “chromosome” consisting of a set of elements called “genes” that hold a set of values for the optimization variables the decision variables. The fitness of each chromosome is determined by evaluating it against an objective function. The chromosome represents a feasible solution for the problem under study. The length of the chromosome equals the number of variables. A Gene value is real coding actual values. The concept of GA is based on the initial selection of a relatively small population. Each individual in the population represents a possible solution in the parameter space. The fitness of each individual is determined by the value of the objective function calculated from a set of parameters. The natural evolutional processes of reproduction selection crossover and mutation are applied using probability rules to evolve the new and better generations. The probabilistic rules allow some less fit individuals to survive. The objective of this study is to maximize the benefit function Z with respect to pumping rate to obtain optimal location of wells Q j as design variable where: Max 2 in which: N w is the total number of pumping wells and P j is penalty. Moreover the management objectives must be achieved within a set of constraints. The constraints may be decision variables. The objective function is subjected to the following constrains: 3.3.1 Objective function constraints 3.3.1.1 Pumping constraint The pumping rates at potential pumping wells in the water demand are constrained to values between some minimum Q j min and maximum Q j max permissible pumping rates as the following: Q j min ≤ Q j ≤ Q j max j ……..N w 3 In the GA simulation this constraint can be easily satisfied by restricting the population space of the design variables to be within the proposed limits. Hence no special treatment is needed for this constraint. 3.3.1.2 Drawdown constraint This constraint normally means to protect the ecosystem by avoiding excessive drawdown. In this work the drawdown constraints were formulated to avoid mining and formulated as follows:-

slide 9:

International Journal of Engineering Research Science IJOER ISSN: 2395-6992 Vol-3 Issue-7 July- 2017 Page | 9 1 w N ji j rd    1 w N jD j QQ    all i D D  at arbitrary point 1 Drawdown constraint r w1 + r w2 + r w3 + r w4 + r w5 +………+ r w68 at arbitrary point 2 Drawdown constraint r w1 + r w2 + r w3 + r w4 + r w5 +………+ r w68 at arbitrary point 3 Drawdown constraint r w1 + r w2 + r w3 + r w4 + r w5 +………+ r w68 at arbitrary point N c Drawdown constraint r w1 + r w2 + r w3 + r w4 + r w5 +……..+ r w68 Accordingly it postulates the relation 4 In which: r w1 is drawdown at point 1 due to pumping well no 1 r w2 is drawdown at point 1 due to pumping well no 2 r w3 is drawdown at point 1 due to pumping well no 3 N c no of control points rj is the drawdown at control point i caused by a pumping rate from pumping well j value of rj in Table 4 to 6 di is the permissible drawdown at control point i equals to 30 m. 3.3.1.3 Water demand constraint The NSAASA is considered as the sole source of water. This therefore means that the designed optimal pumping strategy must supply at least the minimum water demand. It is formulated as follows: 5 In which Q D is water demand. 3.3.1.4 Distance between wells constraint This constraint obeys the relation: 6 Where D i is actual distance between wells and D all is minimum distance between wells. 3.3.1.5 Location of wells The locations of wells constraint is to be decided by the model itself within a user defined region of the model grid until the optimal location is reached Figure 9. FIG. 9: PLAN VIEW SHOWING FINITE DIFFERENCE GRID WITH SUB-DOMAIN FOR WELL LOCATIONS

slide 10:

International Journal of Engineering Research Science IJOER ISSN: 2395-6992 Vol-3 Issue-7 July- 2017 Page | 10 10 0 ff f    3.4 Optimization Procedure of the Simulation-Optimization model In groundwater management problems there are two sets of variables decision variables and state variables where the decision variable is the pumping and injection rates of wells. Also other decision variables include well locations and the on/off status of a well. By optimization techniques the decision variables can be managed to identify the best combination of them. Moreover hydraulic head is the state variable. The simulation model updates the state variables and the optimization model determines the optimal values for all decision variables. In this work MODFLOW FORTRAN code is used as the simulation of groundwater flow which links with GA optimization OLGA Code. The flowchart for simulation-optimization model is given in Fig. 10. FIGURE 10: FLOWCHART FOR THE SIMULATION/OPTIMIZATION MODEL It must refer that MODFLOW simulation without wells is used to generate the initial head but initial population is generated from GA via random number for each well according to constraints and number of chromosome. 3.4.1 Checking the termination criterion The optimization algorithm continues to be executed by iterating until the given termination criteria are satisfied Fig. 10. Note that different termination criteria can be used to stop the computation. These may be: Stopping the computation after a given number of iterations reaching a specific objective function value no improvement in the objective function value for a specified number of passed iterations or after limited time. In this study the relative error definition is used to check the convergence of iterations. Thus the convergence criterion is defined as 7 Where ε is a tolerance for the convergence of iterations if two consecutive objective values f1 f o satisfy the criterion.

slide 11:

International Journal of Engineering Research Science IJOER ISSN: 2395-6992 Vol-3 Issue-7 July- 2017 Page | 11 3.5 Testing scenarios The testing scenarios include three proposed pumping rates and well locations policies. The first scenario estimates the pumping rates from 75 productive wells with their present locations. The second scenario checks the present locations for these 75 productive wells and the optimality of their pumping rates. The third one proposes water exploitation policy aimed at increasing the present productive wells by 14 wells as a result of the increase in reclamation activities new 25 thousand acres and delineates the optimal locations of these new wells and keeping the present locations of the other 75 present wells besides estimating the optimal pumping rates for all. IV. RESULTS AND DISCUSSIONS The expected interference between cones of depression of the present productive wells due to their configurations around the highway roads of the Abu Simbel area will affect the groundwater productivity of the NSAASA. Accordingly the need to determine optimal well location strategy for the new drilled wells required for future expansion of the project 1.5 million acres national project is essential. In addition applying the constructed OLGA Code for solving the optimization problems related to the optimal pumping rates and the optimal well locations the initial pumping rates are assigned to each well as more than 700 m 3 /day and less than 1000 m 3 /day and then the optimization procedure explained before is used. The GA parameters used in the optimization problem are given in Table 4. The optimal well location and pumping rate are determined based on OLGA Code from one feasible solution to the other according to the calculated residual errors values where they improve quickly in the early iterations and change slowly in subsequent iterations till satisfying the constraints. TABLE 4 THE GA SOLUTION PARAMETERS APPLYING OLGA CODE Continuous GA Parameter Value Population size 200 Mutation ratio 0.006 Type of crossover uniform crossover Crossover probability 0.7 Tolerance for the convergence of iterations 0.01 On the other side the results of the three testing optimal policies applying OLGA Code show the future predictions in the optimal pumping rates after 5 10 15 20 25 30 40 and 50 years. The first scenario investigates the proposed pumping rates from the 75 productive wells penetrating the NSAASA under their current locations. The OLGA simulation results show that the optimal pumping rates under the present conditions of water demand Q D of 60000 m 3 /day range from 57585 to 50144 m 3 /day with corresponding total drawdown ranges from 3.42 to 13.73 m during the simulation period of 50 years Table 5. TABLE 5 OPTIMAL PUMPING RATE FOR FIRST SCENARIO Year 5 10 15 20 25 30 40 50 N 75 75 75 75 75 75 75 75 r m 3.42 5.62 6.78 7.88 8.93 9.96 11.85 13.73 Q min /well 746.8 726.4 708.6 695.6 645.2 629.2 613.8 605.7 Q max /well 1000 1000 1000 1000 1000 1000 1000 1000 Q opt m 3 /day 57585 56033 54511 53021 52292 51568 50852 50144 N is number of operation wells r is maximum drawdown Q min /well is min optimal pumping rate for well Q max /well is max optimal pumping rate for well and Q opt total optimal pumping rate for all wells. Moreover by running the developed OLGA Code for time step of 5 years the predicted head distribution maps of the NSAASA under the current pumping rates at years 2015 2025 2035 and 2060 are shown in Figure 11. It is noticed from Figure 11 that the groundwater levels change from 164.5 masl to 161.3 masl in the center of the cone of depression. Few productive wells in the southwestern part of the model domain show relatively slight drawdown amounts to 13 m at East El- Emaratiah Company well field. This is mainly attributed to the effect of increasing the thickness of water bearing formation towards the northwestern and western parts. In addition one large cone of depression will develop in the cultivated areas.

slide 12:

International Journal of Engineering Research Science IJOER ISSN: 2395-6992 Vol-3 Issue-7 July- 2017 Page | 12 This cone is in the central part of Abu Simbel-Tushki road and may attribute to the present extraction rates from the reclaimed areas. FIGURE 11: PREDICTED HEAD DISTRIBUTION MAP OF THE NSAASA FOR OPTIMAL PUMPING RATES APPLYING 1ST SCENARIO a at 2015 b at 2025 c at 2035 and d at 2060 Otherwise the second scenario seems to be theoretical more than practical since it studies the optimal locations of already 75 present productive wells under the present constraint of water demand Q D of 120000 m 3 /day and predict their optimal pumping rates. The GA parameters used in this optimization case are given in Table 6. The initial pumping rates are assigned to each well as less than 1500 m 3 /day and the optimal pumping rates after 5 10 15 20 25 30 40 and 50 years are

slide 13:

International Journal of Engineering Research Science IJOER ISSN: 2395-6992 Vol-3 Issue-7 July- 2017 Page | 13 predicted Table 6. It is noticed from Table 6 that the number of the total operating wells N is still constant 75 wells and the maximum drawdown r ranges from 2.22 to 12.13 m while Q min /well begins with 1460.6 and ends by 1053.2 m 3 /day for the simulation period. The total optimal pumping rate for all wells Q opt ranges between 102038 and 100773 m 3 /day Table 6. TABLE 6 OPTIMAL PUMPING RATE FOR THE 2ND SCENARIO APPLYING OLGA CODE Year 5 10 15 20 25 30 40 50 N 75 75 75 75 75 75 75 75 r m 2.22 4.32 5.43 6.63 7.73 8.64 10.53 12.13 Q min /well 1460.6 1369.5 1301.6 1286.4 1201.3 1183.2 1102.4 1053.2 Q max /well 1500 1500 1500 1500 1500 1500 1500 1500 Q opt m 3 /day 102038 101797 101561 101330 101103 100992 100882 100773 In the same side the results of the OLGA Code based on this scenario show that the predicted optimal location of the productive wells differ more or less from the present location Figure 12. Although the proposed OLGA approach successfully identified all the well locations and pumping rates of the present numerical study identification of all the well locations and pumping rates may not be feasible for real-world applications since the number of pumping wells is usually very high. Instead the proposed procedure may be applied to the suspected illegal pumping areas as noticed in the field trips. As an extension of this study the performance of the proposed OLGA Code should be tested on a real field system in a future study. FIGURE 12: OPTIMAL LOCATION OF PRESENT WELLS 2ND SCENARIO Otherwise the predicted head distribution maps of the NSAASA under the 2 nd scenario Figure 13 show groundwater levels change from 165.5 masl to 162.5 masl in the center of the cone of depression. It is noticed from Figure 13 that the normal decline in the groundwater levels 3 m may attribute to the small thickness and low hydraulic conductivity of the NSAASA in these localities. Moreover the cone of depression will appear in the cultivated areas during the simulation time with diameter of 2.5 to 4 km approximately. While the southern and western cultivated parts of the model domain does not be

slide 14:

International Journal of Engineering Research Science IJOER ISSN: 2395-6992 Vol-3 Issue-7 July- 2017 Page | 14 affected with this scenario. This may attribute to the effect of the great aquifer thickness and the presence of thin clay layers in the southern and western parts rather than the geologic structure impact. Accordingly this reflects the low potentiality of the northern cultivated parts for positive response to this scenario. FIGURE 13: PREDICTED HEAD DISTRIBUTION MAPS OF THE NSAASA FOR OPTIMAL WELL LOCATIONS APPLYING 2ND SCENARIO A AT 2015 B AT 2025 C AT 2035 AND D AT 2060 The third scenario proposes water exploitation policy aimed at increasing the present productive wells by 14 wells as a result of the increase in reclamation activities under the constraints of 500 m as minimum distance between wells and 900000 m 3 /day as water demand Q D and delineates the optimal locations of these new productive wells and keeping the present locations of the other 75 present productive wells besides estimating the optimal pumping rates for all. The GA parameters

slide 15:

International Journal of Engineering Research Science IJOER ISSN: 2395-6992 Vol-3 Issue-7 July- 2017 Page | 15 used in this optimization case are given in Table 7. The OLGA Code runs for time step of 5 years. The predicted optimal location of the new productive wells is checked Figure 14 and the predicted head distribution maps of the NSAASA for the optimal pumping rates at years 2015 2025 2035 and 2060 are shown in Figure 15. TABLE 7 OPTIMAL PUMPING RATE FOR THIRD SCENARIO Year 5 10 15 20 25 30 40 50 N 89 89 89 89 89 89 89 89 r m 4.27 6.47 7.63 8.72 9.77 10.80 12.68 14.56 Q min /well 923.2 913.3 902.8 879.6 823.6 813.2 800 800 Q max /well 1500 1500 1500 1500 1500 1500 1500 1500 Q opt m 3 /day 87187 85635 84114 82623 81894 81171 80454 79747 FIG. 14: OPTIMAL LOCATION OF THE NEW PRODUCTIVE WELLS BLUE COLORED 3RD SCENARIO It is noticed from Table 7 that the maximum drawdown r ranges from 4.27 to 14.56 m while Q min /well begins with 923.2 and ends by 800 m 3 /day in case of number of total operating wells N of 91 productive wells. Moreover Q min /well begins with 923.2 and ends by 800 m 3 /day during the simulation period. The total optimal pumping rate for all wells Q opt ranges between 87187 and 79747 m 3 /day. However it is noticed from the predicted head distribution maps Figure 15 that the groundwater level decreases from 163m asl in the beginning of the simulation period the year 2015 to 158m asl in the end of the simulation period the year 2060. The depression cones related to the first and the second scenarios are mitigated in the northern part of the model domain. Moreover limits are placed on the total optimal pumping rates of the productive groundwater wells beside its optimal locations in the reclaimed areas. Accordingly it is preferred to extend the future reclamation activities parallel to the western side of the Khour Tushki area. Although the results of this scenario provide more or less good degree of confidence in the optimal location of the new productive wells and optimal pumping rates the expected sharp decline in groundwater levels in the concerned area assumes more applied studies. The Q opt in the 3 rd scenario is actually lower than that in the 2 nd scenario because the second scenario studies the optimal locations for these 75 productive wells and their optimal pumping rates. The third one proposes water exploitation policy aimed at increasing the present productive wells by 14 wells. The predicted value of r based on the 3 rd scenario is more or less similar to the results of the other two scenarios although the number of the operating wells is increased by 19. This reflects the great importance to apply the optimal well location concept in any

slide 16:

International Journal of Engineering Research Science IJOER ISSN: 2395-6992 Vol-3 Issue-7 July- 2017 Page | 16 new reclamation projects. As a result of the above discussion of the OLGA Code optimal well location and pumping rates results it can be concluded that the groundwater withdrawal from the NSAASA under this optimization study could be safely conducted. FIGURE 15: PREDICTED HEAD DISTRIBUTION MAPS OF THE NSAASA FOR OPTIMAL WELL LOCATIONS APPLYING 3RD SCENARIO a at 2015 b at 2025 c at 2035 and d at 2060 V. CONCLUSION AND RECOMMENDATION This paper is devoted to propose an assisted methodology to identify representative models for optimal well locations. To do so first a mathematical equations were developed to model the representativeness of a subset of constrains with respect to the full set that characterizes the well location problem. Then an optimization tool was implemented to identify the representative constrains of any problem considering not only the main output variables but also the probability distribution of the attribute-levels of the problem. Accordingly OLGA Code was linked MODFLOW with genetic algorithm technique to establish a simulation optimization groundwater model. This model was applied for NSAASA to develop the optimal location of the groundwater productive wells with maximum pumping rate. Three scenarios were tested to choose the proper

slide 17:

International Journal of Engineering Research Science IJOER ISSN: 2395-6992 Vol-3 Issue-7 July- 2017 Page | 17 water exploitation policy. For the first scenario the predicted groundwater drawdown after simulation period of 50 years ranged from 3.42 to 13.73 m while the corresponding optimal pumping rates ranged from 57585 to 50144 m 3 /day. Under the second proposed scenario the predicted drawdown r and the corresponding total optimal pumping rate for all wells Q opt ranged from 2.22 to 12.13 m and between 102038 and 100773 m 3 /day respectively. Moreover the predicted value of r based on the 3 rd scenario was more or less similar to the results of the other two scenarios although the number of the operating wells was increased by 19. Computational restrictions were common and likely the cause of the systemic lack of algorithm benchmarking for well location optimization. Still some valuable insight can be gained from these results that could guide further investigation. Based on the results of the OLGA model it is highly recommended to choose the new productive well locations in staggered system parallel to the western side of the Khour Tushki area. More applied studies are needed for verification the results of this optimal well location model. Also this study should be extended to contain more test cases and more comprehensive selection of algorithms including interpolation methods ad joint methods and hybrid methods. ACKNOWLEDGEMENTS The data of this work were presented from Desert Research Center Station in Abu Simbel through Development of South Wadi Project Egypt Tushki project which carried out by the team work of DRC. Assistance from project team work and Groundwater Sector Office is gratefully acknowledged. I’m indebted to them especially Prof Himida and Dr El-Sheikh for their help in performing field works and providing the required data. I appreciate the insightful comments and suggestions of the editor and anonymous reviewers which greatly improved this manuscript. REFERENCES 1 Aggour T. A. Korany E. A. Mosaad S. and Kehew A. 2012: Geological conditions and characteristics of the Nubia Sandstone aquifer system and their hydrogeological impacts Tushka area south Western Desert Egypt. Egy. J. Pure Appl. Sci. 2012 027- 037 . 2 Aral M. M. Guan J. B. Maslia M. L. 2001: Identification of contaminant source location and release history in aquifers. Journal of Hydrologic Engineering 6 3 225–234 . 3 Ayvaz MT Karahan H 2008 "Simultaneous parameter identification of a heterogeneous aquifer system using artificial neural networks". Hydrogeology Journal 16: 817–827. 4 Bear J. 1979: Hydraulics of groundwater. McGraw-Hill New York 569 p. 5 Chang L. C. Chu H. J. Hsiao C. T. 2007: Optimal planning of a dynamic pump-treat-inject groundwater remediation system. Journal of Hydrology 342 295–304 . 6 CONOCO 1987: Geological map of Egypt NF 36 NW El Sad El-Ali. Scale 1:500000. The Egyptian General Petroleum Corporation. Conoco Coral . 7 Desert Research Center DRC 2012 "Evaluation of Hydraulic Parameters of Nubia Sandstone Aquifer" unpublished internal report submitted MARA – 65 p. 8 El-Sabri M. A. Sh. El-Sheikh A. E. and El-Osta M. M. 2010: The hydrogeological connection between Lake Nasser and the Nubian Sandstone aquifer in Tushka area South Western Desert Egypt. Egyptian Journal of Aquatic Research 2010. 36 2 245-262. 9 El-Shazly E. M. Abdel Hady M. A. El Ghawaby M. A. Soliman A. B. El-Kassas I. A. Khawaski S. M. El Amin H. El Rakaiby M. M. El Aasy I. E. Abdel Megid A. A. and Mansour S. I. 1980: The geologic map of Egypt the structural lineation map of Egypt and the drainage map of Egypt based on LANDSAT-1 Imagery interpretation and field investigations: Remote Sensing Center Academy of Scientific Research and Technology Cairo: Oklahama state University Stillwater Oklahama and United Nations Environment Program GEMS: 12 maps. Scale 1:100000 . 10 Ferreira da Silva J. Haie N. 2007: Optimal locations of groundwater extractions in coastal aquifers. Water Resources Management 21 8 1299–1311 . 11 Gad M. I. and Khalaf. S. 2013: Application of sharing genetic algorithm for optimization of groundwater management problems in Wadi El-Farigh Egypt. Journal of Applied Water Science. 3:701–716 . 12 Gad M. I. El-Sheikh A. and El-Osta. M. M. 2011: Optimal management for groundwater of Nubian aquifer in El-Dakhla depression Western Desert Egypt. Journal of Water Resources and Environmental Engineering Vol. 314 pp. 393-409 . 13 Guan J. Kentel E. and Aral M. M. 2008: Genetic algorithm for constrained optimization models and its application in groundwater resources management. J. Water Resour Plann. Manag. 134 1 : 64–72. 14 Guan J. Aral M.M. 1999: Optimal remediation with well locations and pumping rates selected as continuous decision variables. Journal of Hydrology 221 20–42 . 15 Hamid R. S. Fatemeh D. Miguel A. M. 2009: Simulation-Optimization Modeling of Conjunctive Use of Surface Water and Groundwater. Water Resour. Manage DOI 10.1007/s11269-009-9533-z . 16 Hill M. C. 1998: Methods and Guidelines for Effective Model Calibration. U.S. Geological Survey Water-Resources Investigations Report 98-4005.

slide 18:

International Journal of Engineering Research Science IJOER ISSN: 2395-6992 Vol-3 Issue-7 July- 2017 Page | 18 17 Hill M. C. 2000:Methods and Guidelines for Effective Model Calibration. Proceedings EWRI ASCE 2000. 18 Huang C. Mayer A.S. 1997: Pump-and-treat optimization using well locations and pumping rates as decision variables. Water Resources Research 33 5 1001–1012 . 19 Karahan H Ayvaz MT "Groundwater parameter estimation by optimization and dual reciprocity finite differences method". J Porous Media 82:211–223 2005 20 Kim J. and Sultan M. 2002: Assessment of the long-term hydrologic impacts of Lake Nasser and related irrigation projects in southwestern Egypt. J. Hydrol. 262 68–83. 21 Liu Q. Zou L. Li 2002: The application of genetic algorithms in groundwater management. World Geology 212:145–149. 22 Mahar P. S. Datta B. 2000: Identification of pollution sources in transient groundwater systems. Water Resources Management 14 209–227 . 23 Mahar P. S. Datta B. 2001: Optimal identification of ground-water pollution sources and parameter estimation. Journal of Water Resources Planning and Management 127 1 20–29 . 24 Mantoglou A. and Kourakos G. 2007: Optimal groundwater remediation under uncertainty using multi-objective optimization. Water Resources Management 21 835–847 . 25 Mckinney D. C. Lin M. 1994: Genetic algorithm solution of groundwater management models. Water Resour Res 306:1897– 1906. 26 Meira L.A.A. Coelho G. P. Santos A.A.S. and Schiozer D.J. 2016. Selection of Representative Models for Decision Analysis Under Uncertainty. Computers Geosciences J.V88 67-82. 27 Moharram S. H. Gad M. I. Saafan T. A. and Khalaf S. 2012: Optimal Groundwater Management Using Genetic Algorithm in El-Farafra Oasis Western Desert Egypt. Water Resour. Manage. 10.1007/s11269-011- 9865-3 . 28 Park C. H. and Aral M. M. 2004: Multi-objective optimization of pumping rates and well placement in coastal aquifers. Journal of Hydrology 290 80–99. 29 Rana T. Khan S. Rahimi M. 2008: Spatio-temporal optimization of agricultural drainage using groundwater models and genetic algorithms: an example from the Murray Irrigation Area Australia. Hydrogeology Journal 16: 1145–1157 . 30 Ruperti N. J. 2002: Estimation of the release history of a contaminant source in 2-D groundwater systems. In: Fourth International Conference on Inverse Problems in Engineering Rio de Janeiro Brazil . 31 Saafan T. A. Moharram S. H. GAD M. I. and Khalaf S.2011: A multi-objective optimization approach to groundwater management using genetic algorithm. Journal of Water Resources and Environmental Engineering Vol. 37 pp. 139-149 August 2011 . 32 Saffi M. Cheddadi A. 2007: Explicit algebraic influence coefficients: a one-dimensional transient aquifer model. Hydrological Sciences Journal-Journal Des Sciences 52 4 763–776 . 33 Singh R. M. Datta B. and Jain A. 2004: Identification of unknown groundwater pollution sources using artificial neural networks. Journal of Water Resources Planning and Management 130 6 506–514 . 34 Sun A. Y. Painter S. L. Wittmeyer G. W. 2006: A robust approach for iterative contaminant source location and release history recovery. Journal of Contaminant Hydrology 88 181–196 . 35 Tung C. P. Chou C. A. 2004: Pattern classification using tabu search to identify the spatial distribution of groundwater pumping. Hydrogeology Journal 12 5 488–496 . 36 Wang W. Ahlfeld D. P. 1994: Optimal groundwater remediation with well location as a decision variable – model development. Water Resources Research 30 5 1605–1618 . 37 Yan T. Wu J. and Xue Y. 2003: A comparison of genetic algorithm and trial-and-error approach in solving the hydrogeologic inverse problem. In: Proceedings of the international Symposium on Water Resources and the Urban Environment. China Environment Science Press Beijing pp 137–141. 38 Yao L. Li J. and Li Z. 2003: Parameter identification of groundwater flow numerical modeling by means of improved genetic algorithm. J Hydraul. Eng. 3412:40–46. 39 Zheng C. M. and Wang P. P. 1999: An integrated global and local optimization approach for remediation system design. Water Resources Research 35 1 137–148 . 40 Zhu X. Wu J. Ye S. Zhao J. and Wu M. 2003: Evaluation of groundwater resource in deep aquifers of the Yangtze Delta south of the Yangtze Delta. In: Proceedings of the international Symposium on Water Resources and the Urban Environment. China Environment Science Press Beijing pp 467–473.

authorStream Live Help