7819 - Laserfiche WebLink

Home Browse Search

1B JOURNAL OF APPLIED AQUACULTURE TABLE 2. Actual and predicted total ammonia nitrogen removal rates at various hydraulic loading rates in two biofilters. The model used to predict ammonia removal rates is described in this paper. The parameter R in the model was deter- mined at the hydraulic loading rate of 44 ml/cmz/minute and used to predict am- monia removal rates at hydraulic loading rates of 44, 22, I1, and 5 ml/cm2/min- ute. Difference Hydraulic Mean Actual Predicted between Flow loading influent ammonia-N ammonia-N actual and rate rate ammonia-N removal rate removal rate predicted (1/min) (ml/cm2/min) fmg/1) Img/h) Img/h) removal rata 10) Gravel + oxygenation biofilter 8 44 1.03 30 (12)a 30 0 4 22 0.86 26 (17) 24 -8 2 11 0.96 23 (11) 25 +9 1 5 1.11 19 18) 26 +37 Fluidized granular carbon • ozonatlon biofilter B 44 ].62 442 (130) 466 +5 2 11 1.33 121 (66) 157 +30 a Standard deviation in parentheses. loading rate of S ml/cm2/minute, which was only 12.5% of the hy- draulic loading rate at which R was determined, the deviation be- tween actual and predicted values increased to 37%. In the granular carbon +ozonatlon biofilter, the predicted ammonia removal rate at a hydraulic loading rate of I1 ml/cm2/minute (using a value of R determined at 44 ml/cmz/minute) was 30% greater than the actual ammonia removal rate (Table 3). Reduced accuracy of prediction over large changes in hydraulic loading rate occurred because Equa- tion 13 compensated for changes in retention time but not for alter- ations in the biomass and activity of the nitrifying bacteria that oc- curred as a result of changes in mass transfer, biofilm shearing, and related factors. The range of hydraulic loading rates over which Equation IZ can generate accurate predictions will vary with the biofilter type. In the