A 4-compartment anaerobic baffled reactor (ABR) was operated for organic wastewater treatment at a constant hydraulic retention time (HRT) of 40 h and temperature of 35°C

with the influent COD increased from 2000 to 8000 mg·L-1 stage by stage. Based on the reactor performance

Andrews model was constructed firstly for substrate degradation in the ABR. And then a prediction model for methane production was constructed by combining the stoichiometry of methane fermentation with the developed substrate degradation model. The results showed that the simulated COD of the calibrated Andrews model was well agreed with the measured data in the ABR. The experimental data were used to calibrate the dynamics parameters. k as the estimated maximum specific substrate degradation rate and Ks as the semi-saturation constant were kept at 2 d-1 and 100 mg COD·L-1

respectively

though the influent COD increased by stages. On the contrary

Ki as the inhibition parameter of volatile fatty acids (VFAs) rose following the increased influent COD. With a COD of 2000

4000

6000 and 8000 mg·L-1 in influent

Ki was counted for 1500

1700

4000 and 6000 mg COD·L-1

respectively. The equivalent COD of VFAs in the first compartment was close to or higher than the estimated Ki

indicating that a serious inhibition of VFAs to methane fermentation had occurred. Meanwhile

the equivalent COD in the last 3 compartments was lower than their Ki and showed a decreasing trend

suggesting that the inhibitory effect of VFAs on methane fermentation became weaker compartment by compartment. The methane production in the 4 compartments was ranged from 1.12—6.42

2.54—8.96

1.24—4.48 and 0.16—0.58 L·d-1

respectively

and the developed model could predict the measured data very well. The calibrated Andrew model and the prediction model for methane production would provide an approach to the design and operation control of ABRs.