Permeable Reactive Barriers in Europe - Permeable Reactive Barrier: Sustainable Groundwater Remediation - page 302

• Homogenization of the flow to provide a more uniform concentra-
tion—achieved by collecting the flow and piping it to the reactor—
that is, separating the flow collector and the reactor. Achieving
uniform flow across the full cross section of a reactor is extremely
difficult at low bed velocities. Homogenization is also helped by
changing the flow direction from horizontal to vertical.
• The vertical orientation allows the use of a greater ratio of flow path
length to the cross-sectional area, thus reducing the potential for short
circuiting—flow concentrating in high flow pathways due to slight het-
erogeneities in the bed. However, it must be accepted that short circuit-
ing remains a serious issue in PRB design because of the very low low
rates (long residence times) required for many PRB reactive materials.
• Estimation of the input flow to PRBs is a major problem—especially
for in-ground reactors as the reactor volume is likely to be constrained.
The author's experience with Belfast and several other PRBs is that the
current modeling techniques, although good at providing flow direc-
tions and groundwater contours, are soft when it comes to predicting
flow rates. The PRB designer has still to accept a wide range of cred-
ible flow rates from the modeler—and design for this range.
• As the cost of a PRB is directly influenced by the flow rate, better
prediction procedures are required. Also, site-pumping test proto-
cols must be refined.
• In-ground reactors are good for sites where the flow is expected to
be low to moderate. A current challenge is to design an in-ground
reactor system for sites where the groundwater flow may be large—
or large for part of a year.
• However, it should not be assumed that better modeling and test-
ing will provide all the answers. The flow through a PRB will vary
seasonally and over longer timescales because of changes to the
groundwater regime resulting from developments around the site
and in the watershed and climatic changes. The risk assessment for
a PRB must consider these factors.
• A PRB may accumulate contaminants as well as destroy them.
Decommissioning must be considered in the design at the outset
(see Carey et al., 2002, for this and for much other advice).
• In Belfast, the reactor was installed in an enlargement in a cement-
bentonite cutoff wall. This was necessary because the source of the
contamination was very close to the site boundary. For later reac-
tors with which the authors have been involved, the reactor has
been placed inside the cutoff wall with only a pipe taken through
the wall. This can be significantly cheaper and ensures that the con-
tamination remains within the funnel.