Masoud Rahimi - CFD Research Center, Chemical Engineering Department, Razi University, Kermanshah, Iran
Saman Khalighi - Department of Biotechnology–Chemical Engineering, Kermanshah Branch, Islamic Azad University,Kermanshah, Iran
Alireza Habibi - CFD Research Center, Chemical Engineering Department, Razi University, Kermanshah, Iran
Sirvan Khalighi - Institute of Systems and Robotics (ISR-UC), Department of Electrical and Computer Engineering, University ofCoimbra, Coimbra, Portugal

In the present study, artificial neural networks were used to predict extent of the chemical oxygen demand (COD) removal and FA degradation rate in bioreactor by Ralstonia eutropha. Initial FA concentration, recycling Substrate flow rate, aeration rate and system’s temperature were used as inputs to the network. Feedforward artificial neural networks with 4-3-2 arrangements, were capable to estimate optimize situation for chemical oxygen demand (COD) remove and FA degradation rate which both of them are the output of the systems. 30 experiments were done and120 data points were collected, so training the ANN with one, three, seven and etc hidden layers using various numbers of neurons were done. The results show that the proposed correlation has good ability for predicting the chemical oxygen demand (COD) remove and FA degradation rate. The result shows satisfactory correlations of R2 = 1.00 and 0.98 in training and testing stages for removalprediction. The proposed neural network models accurately estimate the effects of operational variables in biodegradation of formaldehyde and can be used in order to optimize the process parameters without having to conduct the new experiments in laboratory

Ralstonia eutropha, formaldehyde degradative ability, Artificial Neural Networks, Modeling