Modeling TOC Breakthrough in Granular Activated Carbon Adsorbers

Linear regression techniques were used to develop practical models to predict total organic carbon (TOC) breakthrough in bituminous granular activated carbon (GAC) adsorbers. Models were developed for two field-scale GAC sizes ( $8\times 30$ and $12\times 40\text{mesh}$ ) and two empty bed contact times (EBCTs) (10 and 20 min). Model input parameters include two water quality variables, influent TOC concentration $\left({\text{TOC}}_0\right)$ and pH, that impact performance. The dependent variables for the models were normalized breakthrough time, throughput in bed volumes, to six fractional ( $\text{TOC}/{\text{TOC}}_0=0.2$ , 0.3, 0.4, 0.5, 0.6, and 0.7) and three mass ( $\text{TOC}=1.0$ , 1.5, and 2.0 mg/L) effluent concentrations. Model development was performed using small-scale breakthrough data from 35 different source waters; external model validation was performed with small-scale breakthrough data from 14 source waters; a sensitivity analysis was performed to ensure that the models effectively capture expected breakthrough trend; and a scalability test was performed to verify the models’ ability to predict breakthrough for field-scale GAC adsorbers.