Modelling and Forecasting the Sludge Bulking in Biological Reactors of Wastewater Treatment Plants by Means of Data Mining Methods
The bulking of active sludge in treatment plant bioreactors occurs very often in communal wastewater works what leads to worsening the abilities of sludge sedimentation and the efficiency of works operation. Because of that there is useful and suitable to model and predict the sludge bulking events in order to take some counteractions. In the paper the data mining methods of Support Vector Machines (SVM), Boosted Trees, Random Forests and Multivariate Adaptive Regression Splines (MARS) have been used for modelling and forecasting the sludge bulking events. By the calculation the measurement data series from 4 years concerning the physical and chemical parameters of wastewater flowing into the treatment plant investigated and the technological parameters of the plant bioreactor were used. The calculation results show that the best sludge bulking model containing the best prediction ability has been received by the MARS method and on another side the worst models have been generated by the Random Forests method.
KeywordsMathematical modelling Data mining methods Sewage treatment processes
- 2.Chan, W.T., Koe, L.C.: A knowledge-based framework for the diagnosis of sludge bulking in the activated sludge process. Water Sci. Technol. 23, 847–855 (1991)Google Scholar
- 6.Sezgin, M., Jenkins, D., Parker, D.S.: A unified theory of filamentous activated sludge bulking. J. Water Pollut. Control Fed. 50, 362–381 (1978)Google Scholar
- 12.Łomotowski, J., Szpindor, A.: Nowoczesne systemy oczyszczania ścieków. Wydawnictwo Arkady, Warszawa (2002)Google Scholar
- 18.Osowski, S., Cichocki, A., Siwek, K.: MATLAB w zastosowaniu do obliczeń obwodowych i przetwarzania sygnałów. Oficyna Wydawnicza Politechniki Warszawskiej, Warszawa (2006)Google Scholar