Research On PH Control In Biogas Anaerobic Fermentation Based On Neural Network

Biogas anaerobic fermentation produces biogas for clean energy, by using anaerobic microbe fermentation. The important parameter influencing microbial metabolic process is hydrolyzate pH. However, in the process of anaerobic fermentation, the pH value is nonlinear, time variation and time lag, which can't establish the accurate model. It's hard for traditional PID to control accurately, and it is prone to have super harmonic oscillation, which resulting in low efficiency of methane production. Therefore, The method of rule self-organizing fuzzy control based on neural network to control the change of pH in anaerobic fermentation process is proposed on the whole.Firstly, the pH value change was analyzed by studying the anaerobic fermentation process. And then a simple control model was constructed to determine the control parametersSecondly, using the approximation ability of neural network for controlled process, hydrolysate pH change is identified. After learning and training, the optimal network structure is obtained. Compared with the sample data, it is found that both of the BP and RBF neural network have a good approximation of pH value changes in the process. Comparing them, RBF neural network has a better recognizing ability. Its training time is shorter, and network structure more simple.Finally, the identified network model replaces the actual fermentation process to carry out the fuzzy control. To solve the time lag of pH change,a rule self-organizing module is increased in fuzzy controller, which can fine-tune the fuzzy control rules table to improve the control accuracy. The simulation results show the control effect of Fuzzy control method based on RBF neural network and rule self-organization is great. The accuracy was improved from 6.8± 0.2 to 6.8 ± 0.1.