Biological-filter fuzzy-neural controller design based on COD reduction rate

This paper proposed a novel method for error reinforcement fuzzy-neural network controller design which is only related to two index — COD reduction rate and its variation. The designed controller could generate fuzzy control rules automatically by reinforcement learning in line with the state parameters of a biological-filter wastewater treatment system. The constructed controller has been tested in a bio-filter wastewater treatment experimental system that demonstrates the validity and practicability.