Simulation-based process optimization for surfactant-enhanced aquifer remediation at heterogeneous DNAPL-contaminated sites.

Widespread use of dense non-aqueous phase liquids (DNAPLs) such as TCE and PCE has resulted in contamination of enormous valuable groundwater resources and become high-priority environmental problems. However, experiences from the past decades have demonstrated that DNAPL-contaminated sites were difficult to investigate and challenging to remediate. In this study, a simulation-based process optimization system was developed through integrating a multidimensional simulator, a multivariate statistical tool and an optimization model within a general framework for supporting decisions of surfactant-enhanced aquifer remediation (SEAR). A 3D multiphase and multi-component subsurface model was first provided to simulate SEAR process; dual-response surface models were then established to build a bridge between remediation actions and system performance; a nonlinear optimization model was then formulated for identifying optimal operating conditions for SEAR operations. The results in simulating a typical PCE spill event and the associated SEAR remediation operations in a heterogeneous subsurface indicated that SEAR would be capable of cleaning up the contaminated aquifer with improved efficiencies and cost-effectiveness compared with direct pump-and-treat actions. The regression-analysis results demonstrated that the proposed dual-response surface models were able to predict system responses under given operating conditions. The optimization results demonstrated that the developed simulation-optimization approach was effective in seeking cost-effective SEAR strategies for DNAPL-contaminated sites. With the developed method, optimum operation conditions under various environmental and economic considerations could be compiled into a database that would supports further studies of on-site process-control with injection and extraction rates being the main control variables.