Advanced machine learning application for odor and corrosion control at a water resource recovery facility

Abstract

The objective of this study was to develop a machine learning (ML) application to determine the optimal dosage of sodium hypochlorite (NaOCl) to curtail corrosion and odor by H 2 S in the headworks of a water resource recovery facility (WRRF) without overly consuming volatile fatty acids (VFAs) that are essential for the enhanced biological phosphorus removal. Given the highly diverse datasets available, three subproblems were formulated, and three cascaded ML modules were developed accordingly. The final ML models, chosen based on performance, were able to predict various targeted variables. More specifically, in Module 1, a recurrent neural network (RNN) was designed to predict wastewater characteristics. In Module 2, a random forest (RF) classifier and a support vector machine (SVM) classifier were built with the information from Module 1 along with other datasets to predict the concentrations of VFAs and H 2 S, respectively. Finally, in Module 3, with the information obtained from Module 2, another RF classifier was developed to predict NaOCl dosage to reduce H 2 S but keeping VFAs within the target range. These efforts are relevant and informative for WRRFs that are considering developing Intelligent Water Systems to predict the wastewater characteristics to make operational improvements. Practitioner Points• A recurrent neural network (RNN) using long short-term memory (LSTM) successfully predicted influent wastewater parameters.• A support vector machine classifier predicted hydrogen sulfide (H 2 S) with 97.6% accuracy.• The concentration of VFAs, an important parameter in EBPR, was predicted using a random forest classifier with 93.4% accuracy.• The optimal NaOCl dosage for H 2 S control can be predicted with a random forest classifier using H 2 S, VFAs, and flow.

Artificial Intelligence-Assisted Prediction of Effluent Phosphorus in a Full-Scale Wastewater Treatment Plant with Missing Phosphorus Input and Removal Data

“…Hence, automatic hyperparameter selection (or hyperparameter optimization) is considered an optimal choice, and some common methods include grid search, random search, and model-based optimization such as Bayesian optimization . The grid search is selected in this LSTM work, as it showed great performance in previous LSTM studies. , The list of hyperparameters for grid search consists of the lag size, the number of hidden layers, the number of neurons, the dropout rate, the learning rate, the batch size, and the epoch size to configure and train the LSTM network.…”

“…26 The grid search is selected in this LSTM work, as it showed great performance in previous LSTM studies. 20,27 The list of hyperparameters for grid search consists of the lag size, the number of hidden layers, the number of neurons, the dropout rate, the learning rate, the batch size, and the epoch size to configure and train the LSTM network.…”

“…For example, the LSTM model successfully provides a one-day advance prediction of wastewater characteristics with a mean absolute error (MAE) between 0.02 and 0.06 compared to those of the RF, autoregressive integrated moving average (ARIMA), extreme gradient boosting (XGBoost), and support vector machine (SVM) models. 20 In addition, the LSTM model was capable of predicting phosphorus removal in the form of phosphate via an activated sludge process with an R 2 close to 0.7−0.8. 21 The ML/DL modeling of a full-scale WWTP was challenged by the lack of essential data, including the influent phosphorus concentration and alum dosage.…”

“…The LSTM can suite incomplete data sets because the previous day’s output data can be used as an input feature for prediction. For example, the LSTM model successfully provides a one-day advance prediction of wastewater characteristics with a mean absolute error (MAE) between 0.02 and 0.06 compared to those of the RF, autoregressive integrated moving average (ARIMA), extreme gradient boosting (XGBoost), and support vector machine (SVM) models . In addition, the LSTM model was capable of predicting phosphorus removal in the form of phosphate via an activated sludge process with an R 2 close to 0.7–0.8 …”