Closure to discussion on Optimal in-situ bioremediation design by hybrid genetic algorithm-simulated annealing
Journal of Water Resources Planning and Management
Professor Baveye states that the BIOPLUME II simulator [employed within our simulation/management (S/O) model] insufficiently represents biological clogging (bioclogging). Bioclogging is the change in porosity and hydraulic conductivity of a saturated porous medium attributable to microbial growth. (Baveye et al. 1998). Bioclogging can significantly reduce saturated hydraulic conductivity near in situ bioremediation wells injecting relatively high levels of carbon or energy sources (Vandevivere et al. 1995).
We and others counsel those designing in situ bioremediation systems to consider the possible effect of bioclogging (Dupin and McCarty 2000). We do not reduce aquifer parameters in our example for the following reasons: we do not inject a carbon source, we inject oxygen at a low 8mg∕L concentration, flow velocity near injection wells would probably force the injected oxygen away from the injection wells and further into the aquifer, and the major intent of the paper is to describe a new hybrid heuristic mathematical optimization technique.
One can approximate biofouling effects within simulation/optimization (S/O) even if the flow and transport model does not simulate pore-scale biofouling in detail. To accomplish this, one can specify through model input how those parameters might change with time or injection during a simulation (personal communication with H. Rifai 2005). These specifications are easily done by using any macroscopic simulator for which one has source code. An S/O model employing such a simulator considers the approximate effects of biofouling during its mathematical optimizations.
For a situation meriting more thorough bioclogging simulation, we would employ a different simulator within the S/O model. One candidate is the transport, biochemistry, chemistry, and clogging (TCBC) model (Thullner et al. 2004). To describe bioclogging, TCBC assumes biomass growth in discontinuous colonies and includes multidimensional pore-scale effects.
Other simulators can be easily coupled with the presented optimizer, as we have done previously with other optimizers. Subsurface transport simulators used recently within our S/O models have included ARMOS (ES&T 1991), MT3DMS (Zheng and Wang 1998), and SEAWAT (Langevin et al. 2003). The S/O models have optimized management of multiphase and nonaqueous phase transport (Cooper et al. 1998), dissolved phase transport (Peralta et al. 2003), and density-dependent salt-water intrusion (Peralta et al. 2004).
In conclusion, within an S/O model, one should employ a simulator suitable for the situation. Some in situ bioremediation situations merit more detailed bioclogging simulators than other situations. Design optimization capabilities will advance as bioremediation simulators improve.
Shieh, H. J. and R. C. Peralta. 2006. Closure to discussion on Optimal in-situ bioremediation design by hybrid genetic algorithm-simulated annealing. ASCE Journal of Water Resources Planning and Management. Apr-May 2006. p 128.