Prediction of Hydrogen Production Using Artificial Neural Network

Biohydrogen production from starch wastewater industry via up-flow anaerobic staged reactor (UASR) was investigated. The reactor was operated at a hydraulic retention time (HRT) of 0.28 d, and different food to micro-organisms ratios (F/M) of 0.5, 0.9, 1.4, 1.9 and 2.8 g-COD/g-VSS.d. Peak hydrogen production rate (HPR) of 246 mmol-H2/L.d was observed at F/M of 1.4 g-COD/g-VSS.d. Artificial Neural Network (ANN) with a three layers feed-forward back-propagation (3-8-4-1) was developed to predict the fermentation of biohydrogen production. The network used the default Levenberg-Marquardt algorithm for training. Network inputs were organic loading rate (OLR) (gCOD/L.d), pH and volatile suspended solids (VSS) yield (mg-VSS/g-starch). Network output was HPR (mmol-H2/L.d). It is observed that, the output tracks the targets very well for training (R2-value= 0.945), validation (R2-value=0.652) and testing (R2-value=0.791). These values can be equivalent to a total response of R2-value= 0.849. In this case, the network response is acceptable, and simulation can be used for entering new inputs.