Time-varying optimal in-situ bioremediation design of groundwater using coupled meshless methods and hybrid metaheuristic algorithm

Abstract

In-situ bioremediation is a cost-effective aquifer restoration technique based on microbial degradation of organic groundwater contaminants into harmless substances. Simulation Optimization (SO) models aids in the effective design of field bioremediation process by determining optimal rates and locations of pumping and injection wells. However, most of the previous studies rely on grid/mesh-based numerical simulators and are focused on identification of rates at fixed well locations, despite the locations being crucial for design. Also, the applicability to highly heterogeneous aquifers is a challenge as yet. In this study, two novel models are proposed for identification of time varying pumping and injection rates and optimal well locations with the objective of minimizing bioremediation cost and are valid for highly heterogeneous aquifers. The Meshless Local Petrov Galerkin (MLPG) and Meshless Weak Strong (MWS) methods are selected as simulators, which have advantages of being truly meshless, and facilitating better adaptive analysis and exploration of locations than existing mesh/grid-based models. In this study, a new Hybrid Differential Evolution and Whale Optimization Algorithm (HDEWOA) is proposed as optimizer and is particularly tailored for groundwater remediation problem. The MLPG-HDEWOA and MWS-HDEWOA are verified for a hypothetical rectangular aquifer and the simulations are in excellent agreement with widely used RT3D models and provide better bioremediation designs with lower costs than existing models available. Further, the efficacy of models is demonstrated by successfully applying to a highly heterogenous field type aquifer with nodal transmissivity variation. Thus, the proposed SO models can be reliably extended for better bioremediation design of field aquifer systems.