Optimization Model for the Design of Multi-layered Permeable Reactive Barriers

Abstract

Permeable reactive barriers (PRBs) are employed as in situ groundwater remediation technology. The installation of PRBs is usually a major investment, where one of the biggest cost drivers are material costs. PRBs are barriers against contaminants moving under the natural gradient, however not against groundwater contaminants. The most common construction of a PRB is a single barrier, but in the case of contaminant mixtures a multi-layered construction, i.e. a combination of different reactive materials and removal processes, is required. The most important parameters for PRB design are dimensions. The barrier must be long enough to treat the entire width of the plume (dimension perpendicular to groundwater flow) and should extend to and be keyed into an impermeable layer. The problem is to determine the optimal thickness of a PRB, which should provide a residence time appropriate for reducing the concentration of contaminants to the desired effluent concentration. In PRBs, design is accomplished using numerical methods or simulators, which are useful to predict the scenarios and evaluate the resulting groundwater flow systems to specific site conditions. On the other hand, numerical methods are complicated and may have significant errors if the discretization is too coarse or is incorrectly aligned. This paper deals with a simple, conceptual model of a one-approach optimization method for multi-layered PRB design. Based on literature and laboratory test results (residence time, density and hydraulic coefficient), a selection of layers of reactive materials was determined. Considering the lowest cost of the reactive materials, the required thicknesses of activated carbon, zeolite and zero valent iron were calculated using two different algorithms. The simple model may be used for preliminary barrier design and cost calculations. Using the optimization model in a preliminary design stage, it is possible to reject the PRB concept and avoid losing time for the complicated analysis.