Detail Abstract

Prediction of WWTP Influent and Effluent Characteristics using Machine Learning

Abstract

The operation of a wastewater treatment plant (WWTP) is a complex task which requires to consider several aspects: adapting to always changing influent composition and volume, ensuring treated effluents quality complies with local regulations, ensuring dissolved oxygen levels in biological reaction tanks are sufficient to avoid anoxic conditions etc. all of it while minimizing usage of chemical and power consumption. The traditional way of managing WWTPs consists in having employees on the field measure various parameters and make decisions based on their judgment and experience which holds various concerns such as the low frequency of data, errors in measurement and difficulty to analyze historical data to propose optimal solutions.</br> There is a need to develop technologies which would allow forecasting both influent and effluent flow as well as main water quality parameters (such as BOD, COD, pH, TN, TP etc.). In this study, forecast of WWTP influent and effluent characteristics was carried out using programs based on ARIMA, ARIMAX and Neural Network algorithms. Influent and effluent flow and quality data was obtained from currently operating WWTP and analyzed to identify temporal trends (daily, weekly, seasonal) as well as correlation between several variables. Then the performance of each one of the three methods was compared: ARIMA studies past variations of a single variable to forecast future values, ARIMAX includes exogenous variables correlated with the variable to forecast as predictors and Neural Network also includes exogenous variables while using a hidden layer of neurons to perform predictions. Each method performance is analyzed using several indices (Pearson Correlation Coefficient, Root Mean-Squared Error or Mean Average Percentage Error) and global results are discussed.