Abstract
Understanding the mechanisms of pollutant removal in Wastewater Treatment Plants (WWTPs) is crucial for controlling effluent quality efficiently. However, the numerous treatment units, operational factors, and the underlying interactions between these units and factors usually obfuscate the comprehensive and precise understanding of the processes. We have previously proposed a machine learning (ML) framework to uncover complex cause-and-effect relationships in WWTPs. However, only one interpretable ML model, Random forest (RF), was studied and the interpretation method was not granular enough to reveal very detailed relationships between operational factors and effluent parameters. Thus, in this paper, we present an upgraded framework involving three interpretable tree-based models (RF, XGboost and LightGBM), three metrics (R2, Root mean squared error (RMSE), and Mean absolute error (MAE)) and a more advanced interpretation system SHapley Additive exPlanations (SHAP). Details of the framework are provided along with a demonstration of its practical applicability based on a case study of the Umeå WWTP in Sweden. Results show that, for both labels TSSe (Total suspended solids in effluent) and PO4e (Phosphate in effluent), the XGBoost models are optimal whereas the RF models are the least optimal, due to overfitting and polarized fitting. This study has yielded multiple new and significant findings with respect to the control of TSSe and PO4e in the Umeå WWTP and other similarly configured WWTPs. Additionally, this study has produced two important generic findings relating to ML applications for WWTPs (or even other process industries) in terms of cause-and-effect investigations. First, the model comparison should be carried out from multiple perspectives to ensure that underlying details are fully revealed and examined. Second, using a precise, robust, and granular (feature attribution available for individual instances) explanation method can bring extra insight into both model comparison and model interpretation. SHAP is recommended as we found it to be of great value in this study.
Highlights
Wastewater Treatment Plants (WWTPs) are systems used to remove various pollutants from collected wastewater to ensure the effluent discharged into the water cycle complies with the regulations and has minimal influence on the environment (Russell, 2019)
Gradient boosting decision tree (GBDT) (Friedman, 2001) is an ensemble method in machine learning where multiple weak learners are combined to form a single strong learner. It is different from bagging methods, and is characterized by the sequential and iterative learning process in which the current regression tree is trained using the residuals from the previous tree
In the tuning process of hyperparameters, the models’ generalization performance i.e. to prevent overfitting of the models was prioritized. This was to guarantee that the results of the SHapley Additive exPlanations (SHAP) interpretation carried out on the training data were applicable to future unknown data
Summary
Wastewater Treatment Plants (WWTPs) are systems used to remove various pollutants from collected wastewater to ensure the effluent discharged into the water cycle complies with the regulations and has minimal influence on the environment (Russell, 2019). In a recently published paper (Wang et al, 2021), based on online monitored process data, we proposed a novel machine learning (ML) framework that can be used to uncover the precise relationship between operational factors and effluent quality In this framework, a Deep Neural Network (DNN) model (Sugiyama, 2019) was used to validate whether the Random Forest (RF) model (Breiman, 2001, 2002) captured sufficient variance to support the further RF model interpre tation – Variable Importance Measure (VIM) analysis and Partial Dependence Plot (PDP) analysis (Friedman, 2001). VIM was carried out to identify the most influential operational factors on effluent quality, and PDP was carried out to investigate how those influential factors affect effluent quality This framework, and its case study on the local Umeå WWTP in Sweden, helped in the development of a more advanced control strategy to optimize the usage of chemicals and energy without compromising effluent quality. The expansion of tree-based models and the adoption of the SHAP system are justified both theoretically and through the case study details to serve as a reference for the studies with the intention of understanding WWTP processes through ML implementation
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
