2019 journal article

Stochastic cost-optimization and risk assessment of in situ chemical oxidation for dense non-aqueous phase liquid (DNAPL) source remediation

STOCHASTIC ENVIRONMENTAL RESEARCH AND RISK ASSESSMENT, 33(1), 73–89.

By: U. Kim*, J. Parker * & R. Borden n

co-author countries: United States of America πŸ‡ΊπŸ‡Έ
author keywords: Stochastic optimization; In situ chemical oxidation; Risk assessment; DNAPL source remediation; Uncertainty analysis
Source: Web Of Science
Added: February 18, 2019

This study involved development of a computer program to determine optimal design variables for in situ chemical oxidation (ISCO) of dense nonaqueous phase liquid (DNAPL) sites to meet site-wide remediation objectives with minimum life-cycle remediation cost while taking uncertainty in site characterization data and model predictions into consideration. A physically-based ISCO performance model computes field-scale DNAPL dissolution, instantaneous reaction of oxidant with contaminant and with readily oxidizable natural oxidant demand (NOD), second-order kinetic reactions for slowly oxidizable NOD, and time to reach ISCO termination criteria. Remediation cost is computed by coupling the performance model with a cost module. ISCO termination protocols are implemented that allow different treatment subregions (e.g., zones with different estimated contaminant concentrations) to be terminated independently based on statistical criteria related to confidence limits of contaminant concentrations estimated from soil and/or groundwater sampling data. The ISCO model was implemented in the program called Stochastic Cost Optimization Toolkit, which includes modules for additional remediation technologies that can be implemented serially or in parallel coupled with a dissolved plume model to enable design optimization to meet plume-scale cleanup objectives. This study focuses on optimization of ISCO design to meet specified source zone remediation objectives. ISCO design parameters considered for optimization include oxidant concentration and injection rate, frequency and number of soil or groundwater samples, and cleanup criteria for termination of subregion injection. Sensitivity studies and example applications are presented to demonstrate the benefits of proposed stochastic optimization methodology.