@misc{Studziński_Jan_Prognozowanie_2016, author={Studziński, Jan and Rojek, Izabela and Szeląg, Bartosz}, copyright={Creative Commons Attribution BY 4.0 license}, address={Warszawa}, journal={Raport Badawczy = Research Report}, howpublished={online}, year={2016}, publisher={Instytut Badań Systemowych. Polska Akademia Nauk}, publisher={Systems Research Institute. Polish Academy of Sciences}, language={eng}, abstract={In the present study the Support Vector Method (SVM) is used to analyze the dependency between input variables (quantity and quality of wastewaters at the inflow and operational characteristics of an Activated Sludge Tank (AST)) and a result of predicted mixed liquor suspended solid (MLSS) and substrate loads (FIM). Computations revealed, that the highest errors for MLSS are present if only the load of organie compounds susceptible to chemical degradation was included in the input data and lowest when load of coal, ammoniacal nitrogen, suspension and ASCh's operational characteristics were used as the input. Moreover, it appeared that indices of wastewater quality at the inflow to the treatment plant can be sirnulated on the basis of the measured discharge and temperature of wastewaters and in a result it is possible replacing these measured indices with modeled ones to sirnulated MLSS and FIM. The lowest 'errors of predicted substrate loads, computed on th.e basis of modeled using SVM indices values were obtained for coal, arnmoniacal nitrogen, suspension loads and AST's operational characteristics - sludge temperature and pH and methanol dosage.}, title={Prognozowanie dopływu ścieków surowych do oczyszczalni i ładunku zanieczyszczeń w ściekach za pomocą sieci neuronowych i modeli operatorowych * Modelling mixed liquor suspended solid and substrate load on the basis of wastewater quality in di ces and operational parameters of the bioreactor: data mining approach}, type={Text}, URL={http://www.rcin.org.pl/Content/208751/PDF/RB-2016-50-02.pdf}, }