A nonparametric, nonlinear regression technique was used to develop models to
predict total organic carbon (TOC) breakthrough time for a wide assortment of small- and
field-scale granular activated carbon (GAC) adsorbers. Model development included an
evaluation of twelve independent (and possibly correlated) variables, of which only four
proved to be of statistical importance: influent TOC concentration, pH, empty bed
contact time (EBCT), and field-scale GAC size. The models were cross-validated using
an internal "leave-one-out" technique; their predictive capabilities and scalability were verified using external data. Includes 18 references, tables, figures.
| Edition : | Vol. - No. |
| Number of Pages : | 20 |
| Published : | 06/17/2005 |
