Measurement of ammonia nitrogen concentration after sewage treatment based on HCPS multilayer perceptrons
GAO Ming
CUI Ke
LI Hao
LI Sanyi
Abstract：In the existing prediction model of ammonia nitrogen in sewage treatment effluent, the hidden layer neurons excessively redundant, which wastes resources. A new structure adjustment algorithm( HCPS) based on sensitivity analysis( SA) and mutual information( MI) was proposed. The algorithm redefined the sensitivity formula, adaptively, adjusted the network structure by using sensitivity and mutual information, deleted hidden neurons with low sensitivity, divided hidden neurons with excessive sensitivity, and merged two hidden neurons with excessive mutual information. The results of verification on BSM1, a benchmark simulation platform for sewage treatment, showed that HCPS algorithm could obtain a more compact network structure, and the prediction accuracy of ammonia nitrogen concentration in effluent was higher.
Keywords：sewage treatmentammonia nitrogen concentration predictionsoft measurementHCPSmultilayer perceptronstructural self-organizing
Publication Date：2018-01-01
Online Publishing Date：2025-08-15(First online date of this platform, not the publication date of the document)
Pages：10( 92-100,108 )