Comparison of batch sorption tests, pilot studies, and modeling for estimating GAC bed life

Saint Paul Regional Water Services (SPRWS) in Saint Paul, MN experiences annual taste and odor episodes during the warm summer months. These episodes are attributed primarily to geosmin that is produced by cyanobacteria growing in the chain of lakes used to convey and store the source water pumped from the Mississippi River. Batch experiments, pilot-scale experiments, and model simulations were performed to determine the geosmin removal performance and bed life of a granular activated carbon (GAC) filter-sorber. Using batch adsorption isotherm parameters, the estimated bed life for the GAC filter-sorber ranged from 920 to 1241 days when challenged with a constant concentration of 100 ng/L of geosmin. The estimated bed life obtained using the AdDesignS model and the actual pilot-plant loading history was 594 days. Based on the pilot-scale GAC column data, the actual bed life (>714 days) was much longer than the simulated values because bed life was extended by biological degradation of geosmin. The continuous feeding of high concentrations of geosmin (100-400 ng/L) in the pilot-scale experiments enriched for a robust geosmin-degrading culture that was sustained when the geosmin feed was turned off for 40 days. It is unclear, however, whether a geosmin-degrading culture can be established in a full-scale filter that experiences taste and odor episodes for only 1 or 2 months per year. The results of this research indicate that care must be exercised in the design and interpretation of pilot-scale experiments and model simulations for predicting taste and odor removal in full-scale GAC filter-sorbers. Adsorption and the potential for biological degradation must be considered to estimate GAC bed life for the conditions of intermittent geosmin loading typically experienced by full-scale systems.