Assessment of the reliability of the operation of a sewage treatment plant using Monte Carlo simulation

Abstract

The aim of the study was to model the operation of a wastewater treatment plant using the Monte Carlo method and selected probability distributions of random variables. Pollutant indices in treated wastewater were analysed, such as: biological oxygen demand (BOD5), chemical oxygen demand (CODCr), total suspended solids (TSS), total nitrogen (Ntot), total phosphorus (Ptot). The preliminary analysis of pollution indicators series included the:calculation of descriptive statistics and assessment of biological degradability of wastewater. The consistency of the theoretical distributions with the empirical ones was assessed using Anderson–Darling statistics. The best-fitting statistical distributions were selected using the percent bias criterion. Based on the calculations performed, it was found that the analysed indicators of pollution in treated wastewater were characterised by an average variability of composition for BOD5, CODCr and TSS, and a high variability of composition for Ntot and Ptot. The best fitted distribution was log-normal for BOD5, TSS, Ntot and Ptot and general extreme values for CODCr. The simulation carried out using the Monte-Carlo method confirmed that there may be problems associated with the reduction of nutrients (Ntot and Ptot) the analysed wastewater treatment plant. Results of values obtained of the risk values of negative control of wastewater treatment plant operation for biogenic compounds, different from 1, indicate that the number of exceedances at the outflow may be higher than the acceptable one.