Artificial Neural Network (ANN) Modelling for Biogas Production in Pre-Commercialized Integrated Anaerobic-Aerobic Bioreactors (IAAB)

The use of integrated anaerobic-aerobic bioreactor (IAAB) to treat the Palm Oil Mill Effluent (POME) showed promising results, which successfully overcome limitation a large space that is needed in conventional method. understanding synergism between anaerobic digestion and aerobic process required achieve maximum biogas production COD removal. Hence, this work presents artificial neural network (ANN) predict removal (%), purity methane yield (LCH4/gCODremoved) biochemical oxygen demand (BOD) total suspended solid (TSS) (%) pre-commercialized IAAB located at Negeri Sembilan, Malaysia. MATLAB R2019b was used develop two ANN models. Bayesian regularization backpropagation (BR) best performance among 12 training algorithms. trained models high accuracy (R2 > 0.997) demonstrated good alignment with industrial data obtained from over 6-month period. developed model subsequently create optimal operating conditions maximize output parameters. improved by 33.9% (from 68.7% 92%), while 13.4% 0.23 LCH4/gCODremoved 0.26 LCH4/gCODremoved). Sensitivity analysis shows inlet most influential input parameters affect yield, COD, BOD TSS removals, for process, affected mixed liquor solids (MLSS). can be utilized as decision support system (DSS) operators behavior solve problems instability inconsistent process. This utmost importance successful commercialization technology. Additional such mixing time, reaction nutrients (ammonium nitrogen phosphorus) concentration microorganisms could considered improvement model.