Modeling of a paper-making wastewater treatment process using a fuzzy neural network |
| |
Authors: | Mingzhi Huang Yan Wang Yongwen Ma Huiping Zhang Hongbin Liu Zhanzhan Hu ChangKyoo Yoo |
| |
Affiliation: | 1.State Key Laboratory of Pulp and Paper Engineering,South China University of Technology,Guangzhou,China;2.School of Chemistry and Chemical Engineering,South China University of Technology,Guangzhou,China;3.College of Environmental Science and Engineering,South China University of Technology,Guangzhou,China;4.The Key Lab of Pollution Control and Ecosystem Restoration in Industry Clusters, Ministry of Education,South China University of Technology,Guangzhou,China;5.School of Business Administration,South China University of Technology,Guangzhou,China;6.Department of Environmental Science and Engineering, Center for Environmental Studies,Kyung Hee University,Gyeonggi-do,Korea |
| |
Abstract: | An intelligent system that includes a predictive model and a control was developed to predict and control the performance
of a wastewater treatment plant. The predictive model was based on fuzzy C-means clustering, fuzzy inference and neural networks.
Fuzzy C-means clustering was used to identify model’s architecture, extract and optimize fuzzy rule. When predicting, MAPE
was 4.7582% and R was 0.8535. The simulative results indicate that the learning ability and generalization of the model was
good, and it can achieve a good predication of effluent COD. The control model was based on a fuzzy neural network model,
taking into account the difference between the predicted value of COD and the setpoint. When simulating, R was 0.9164, MAPE
was 5.273%, and RMSE was 0.0808, which showed that the FNN control model can effectively change the additive dosages. The
control of a paper-making wastewater treatment process in the laboratory using the developed predictive control model and
MCGS (monitor and control generated system) software shows the dosage was computed accurately to make the effluent COD remained
at the setpoint, when the influent COD value or inflow flowrate was changed. The results indicate that reasonable forecasting
and control performances were achieved through the developed system; the maximum error was only 3.67%, and the average relative
error was 2%. |
| |
Keywords: | |
本文献已被 SpringerLink 等数据库收录! |
|