Advanced controlling of anaerobic digestion by means of hierarchical neural networks

In this work several feed-forward back propagation neural networks (FFBP) were trained in order to model, and subsequently control, methane production in anaerobic digesters. To produce data for the training of the neural nets, four anaerobic continuous stirred tank reactors were operated in steady-state conditions at organic loading rates (Br) of about 2 kg m−3 d−1 chemical oxygen demand, and disturbed by pulse-like increase of the organic loading rate. For the pulses additional carbon sources like flour, sucrose, 1,2-diethylen glycol or vegetable oil were added to the basic feed, which consisted of surplus and primary sludge of a local waste-water treatment plant, to increase the chemical oxygen demand. Measured parameters were: gas composition, methane production rate, volatile fatty acid concentration, pH, redox potential, volatile suspended solids and chemical oxygen demand of feed and effluent. A hierarchical system of nets was developed and embedded in a decision support system to find out which is the best feeding profile for the next time steps in advance. A 3-3-1 FFBP simulated the pH with a regression coefficient of 0.82. A 9-3-3 FFBP simulated the volatile fatty acid concentration in the sludge with a regression coefficient of 0.86. And a 9-3-2 FFBP simulated the gas production and gas composition with a regression coefficient of 0.90 and 0.80, respectively. A lab-scale anaerobic continuous stirred tank reactor controlled by this tool was able to maintain a methane concentration of about 60% at a rather high gas production rate of between 5 and 5.6 m3 m−3 d−1.