Enhanced Simulation-Optimization Approach Using Surrogate Modeling for Solving Inverse Problems

Abstract

This study investigates and discusses groundwater system characterization problem utilizing surrogate modeling. In this inverse problem, the contaminant signals at monitoring wells are recorded to recreate the pollution profiles. In this study, simulation-optimization approach is a technique utilized to solve inverse problems by formulating them as an optimization model, where evolutionary computation algorithms are used to perform the search. In this approach, the partial differential equations (PDE) groundwater transport simulation model is solved iteratively during the evolutionary search, which in general can be computationally expensive since thousands of simulation model evaluations will be evaluated. To overcome this limitation, the simulation model is replaced by a surrogate model, which is computationally much faster than the simulation model and yet is relatively accurate. Artificial neural networks (ANN) is used to construct surrogate models that provide acceptable accuracy performances. The ANN surrogate model, which replaces the PDE groundwater transport simulation model, is then coupled with a genetic algorithm (GA) search procedure to solve the source identification problem. The results will present the quality solution of the ANN surrogate model versus the groundwater simulation model, the solution of the inverse problem for different experiment scenarios and finally a timing study analysis conducted to measure the surrogate model performance.

Residual-based adaptivity for two-phase flow simulation in porous media using Physics-informed Neural Networks

“…A successful approach for solving inverse problems relies on building surrogate parametric models for fast exploration of the state space of parameters [18,19]. These meta-models are specifically designed to be an optimal trade-off between accuracy and computational cost and can be effectively used for inverse problems when the minimization of the objective function requires multiple calls to the flow solver to perform parametric sweeps.…”

Contaminant source localization via Bayesian global optimization

“…In this group of methods, both the flow field and the contaminant plume are assumed to be perfectly known. Methods using forward solvers are based on an inverse problem formulation (Aral et al, 2001;Yeh et al, 2007;Mirghani et al, 2012), where the source location and release history are inferred from concentration samples. Parameter sets are proposed and used as inputs in a forward solver to simulate concentration breakthrough curves at the sample locations; when the mismatch between the simulated concentrations and the observed ones is within an acceptable level of error, the proposed model is accepted as a solution.…”

Application of Mixed-Integer Nonlinear Optimization Programming Based on Ensemble Surrogate Model for Dense Nonaqueous Phase Liquid Source Identification in Groundwater

“…G roundwater contamination source identification (GCSI), including number, location, and release history (Atmadja and Bagtzoglou, 2001;Sun et al, 2006;Sun, 2009), is critical for taking effective measures to protect groundwater resources, assess risks, mitigate disasters, and design remediation strategies (Mirghani et al, 2012;Om and Bithin, 2013).…”

“…The artificial neural network method is most widely used as a surrogate model for numerical simulation of GCSI (Singh et al, 2004;Mirghani et al, 2012;Singh, 2014, 2015). However, it suffers from instability and overfitting problems that are difficult to solve.…”