Kinetic study of biological hydrogen production by anaerobic fermentation

Bio-hydrogen production from thin stillage using conventional and acclimatized anaerobic digester sludge

“…This trend is similar to that observed by Jinming et al [11] who used food waste as the substrate and anaerobic digester sludge as the seed, where the hydrogen yield increased slowly to a maximum of 39 mL H 2 /gVS at S 0 /X 0 ratio of 6 gVS feed /gVS seed prior to decreasing to almost zero at S 0 /X 0 ratio of 8 gVS feed / gVS seed and higher. In addition, in another study [12], the same trend was observed in batches using seed sludge from a local anaerobic digester and food waste as the substrate, with a maximum yield of 101 mL H 2 /gCOD at S 0 /X 0 ratio of 7.68 gCOD/gVSS. The differences in the optimum S 0 /X 0 ratios in the literature can be attributed to the differences in the waste type and characteristics as well as the anaerobic digester sludges.…”

Application of artificial neural networks for modeling of biohydrogen production

“…r e f e r e n c e s 5 e Experimental and predicted hydrogen production profile using data from [24,26,28]. Fig.…”

“…The ANN model was then used to estimate the temporal hydrogen evolution for three new data sets adopted from Chen et al [24], Nasr et al [26], and Wang and Wan [28] that were not used in the training process. Chen et al [24] investigated biohydrogen production from sucrose in batch studies using ADS at 36 C and an initial pH of 5.5.…”

“…Chen et al [24] investigated biohydrogen production from sucrose in batch studies using ADS at 36 C and an initial pH of 5.5. Nasr et al [26] investigated biohydrogen production from thin stillage as the substrate using ADS as the seed microflora at 37 C and an initial pH of 5.5.…”

“…4 shows the correlation between the predicted and experimental data points from the aforementioned sets of data, where a correlation coefficient of 0.965 and an MAE of 11.2 mL were obtained. Average percentage error (APE), defined as the summation of the absolute difference between the experimental and predicted values divided by the experimental values, averaged over the number of data points were 1.4%, 9.6%, and 8.5% for the data sets adopted from Nasr et al [26], Chen et al [24], and Wang and Wan [28], respectively. Fig.…”