Abstract
Aiming at the difficulty of real-time and accurate measurement of biochemical oxygen demand (BOD) in effluent of wastewater treatment process, a design method based on modular neural network is proposed to predict BOD in effluent. Firstly, the BOD sequence is decomposed by Complementary Ensemble Empirical Mode Decomposition (CEEMD) algorithm, Fuzzy Entropy (FE) is used to quantify the complexity of each subsequence. Then, using density clustering to determine the adaptive RBF sub-network structure, and Error Feedback and Sensitivity Analysis (SA) could dynamically adjust the number of neurons. On this basis, the network width is adjusted by clustering algorithm. Experiments verify that this method has a more concise network structure and higher prediction accuracy.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
