1 1 An Optimal Strategy For Groundwater Remediation By Coupling Groundwater Modeling System (GMS) and Particle Swarm Optimization (PSO) Partha Majumder 1 and T I Eldho 2 1 Research Scholar, Department of Civil Engineering, Indian Institute of Technology Bombay, Mumbai 400076, India. E-mail: parthamajp@gmail.com 2 Professor, Department of Civil Engineering, Indian Institute of Technology Bombay, Mumbai 400076 E-mail: eldho@civil.iitb.ac.in Abstract: Groundwater contamination is a major problem in many parts of the world. Once contaminated, remediation of groundwater becomes a tedious, time consuming and expensive process. For the groundwater remediation, pump and treat (PAT) is one of the commonly used technique, in which the contaminated groundwater is pumped, treated and put back into the aquifer system or other sources. Simulation-Optimization (S/O) models are very useful in the appropriate design of an effective PAT system. In the S/O models, while simulation model helps to predict the spatial and temporal variation of the contamination plume, optimization models are used to minimize the cost of pumping. Due to non-linear, non-analytical nature of the groundwater flow and transport processes, the numerical methods are a must to solve the problem. In this study, a groundwater management model has been developed for groundwater remediation using PAT by coupling user friendly commercial software Groundwater Modeling System (GMS) and Particle swarm optimization (PSO). Groundwater flow and transport processes has been simulated by using MODFLOW and MT3DMS package available in GMS. Thereafter a MATLAB function has been developed to execute MODFLOW and MT3DMS in MATLAB environment. Further, MODFLOW and MT3DMS is coupled with particle swarm optimization (PSO) to develop an optimal groundwater remediation strategy. The proposed methodology is applied to a hypothetical homogeneous confined aquifer for the determination of minimum remediation cost. MODFLOW and MT3DMS were also coupled with Genetic Algorithm (GA) and results are compared with the result of PSO based model. It has been observed that the PSO based model gives much better result than GA based model with less computational time. Although, the problem considered is hypothetical, the proposed methodology can be very easily implemented for real field conditions. Index terms - Groundwater remediation, Pump and treat (PAT), Simulation – optimization, Groundwater Modeling System (GMS), MODFLOW, MT3DMS, Particle Swarm Optimization (PSO), I. INTRODUCTION: Groundwater is the most significant source of water for a variety of uses, like industrial, irrigation, domestic and drinking purposes. However, the groundwater is extremely susceptible to contamination and it is one of the most important health related environmental problems. Some of the major sources of groundwater contamination are: fuel storage tank leakage, waste disposal sites, landfills, pesticides, fertilizers etc. [1]. If groundwater contamination happens, then it is necessary to remediate the aquifer or at least contain the contaminants, in order to prevent the further transport of contaminants in river, lakes or public water supply wells. The remediation of the contaminated groundwater is a complicated and expensive and long term process. The pump and treat method (PAT) is one of the most commonly used method for the remediation of contaminated groundwater. In a pump and treat method, a number of pumping wells are installed to extract the contaminated groundwater. The contaminated water is then treated above the ground and thereafter the water is either injected back into the aquifer or may be used for agriculture, forestry and household work [2]. In PAT method, generally a simulation model, which simulates the governing groundwater flow and solute transport equations, is coupled with an optimization technique for developing a strategy of aquifer remediation process.