Soft Measurement of Effluent Ammonia Nitrogen Based on IABC-MFFNN

Abstract

In wastewater treatment processes (WWTPs), it is difficult to accurately measure effluent ammonia nitrogen. a soft measurement method of effluent ammonia nitrogen is proposed based on improved artificial bee colony of multi-feedback fuzzy neural network (IABC-MFFNN). Firstly, multiple feedback links are added to the fuzzy neural network (FNN). Due to existence of internal feedback and external feedback, the system can perform self-adaptive adjustments based on previous status information and output signals, which can fully reflect the dynamic information. Secondly, an improved artificial bee colony (IABC) is proposed as the parameter optimization method of the network. The multi-strategy selection mechanism was used to design different search strategies for each subgroup, and the double index dynamic subgroup mechanism was used to adjust the number of food sources for each ordinary subgroup, which improved the accuracy of the soft measurement method. Simulation experiment results show that the proposed method has higher prediction accuracy compared with other methods.