Evolutionary Prediction of Biohydrogen Production by Dark Fermentation

The present work is a study of the performance and effect of operational parameters on biohydrogen production from palm oil mill effluent by dark fermentation in batch mode. The process parameters examined are pH (5, 5.5, and 6), temperature (30, 35, and 40 °C), substrate concentration (5000, 12 500, and 20 000 mg L−1) and inoculum volume (20, 25, and 30 mL). The inoculum concentration prepared was 10 000 mg L−1 volatile suspended solids. The experiments were designed by response surface methodology (RSM). The highest chemical oxygen demand (COD) removal, hydrogen percentage (H2%) and hydrogen yield (HY) obtained were 58.3%, 80%, and 4.83 mol H2/mole of COD consumed, respectively. Based on the experimental data obtained with the RSM design, gene expression programming (GEP) was developed to predict the COD removal, hydrogen production, and hydrogen yield as process responses. The projected models were assessed based on the correlation coefficient (R2), root mean square error, mean absolute relative error, scatter index, and BIAS. The results demonstrate that the GEP model outperformed the RSM model and was superior in predicting the response variables. Partial derivative sensitivity analysis was also employed to assess the effect of each variable on COD%, H2%, and HY prediction. The prediction uncertainty for COD%, H2%, and HY was quantified, and the results were ±0.11, ±0.17, and 0.015, respectively. According to the results, the GEP model is more efficient than the RSM model in predicting the experimental data for biological hydrogen production in the dark fermentation process. The influence of the most significant operational parameters on biological hydrogen production from palm oil mill effluent by dark fermentation is studied. Response surface methodology and a gene expression programming method are applied as a robust tool for modeling and indicating individual relationships. The results are compared to identify the best estimation of the operational variables affecting biological hydrogen production.