International Journal of Environmental Science and Development

Abstract—Optimization and control of waste water treatment plants (WWTP) is an ongoing effort to make the process more efficient and cost-effective. As found in literature, data mining models such as neural networks have been applied to simulate and model various aspects of the plant such as performance, quality parameters and process parameters. In this paper, we introduce bagging model, an ensemble data mining model, to predict the performance of the WWTP. Ensemble models have been shown to stabilize the base classifier used and avoid overfitting the data. Bagging was used to predict the performance of individual units (primary settler and secondary settler) and the global plant performance. The predicted performance of individual units was also used as inputs to predict the global performance thereby enabling good process control via predictive data models. Upon application to the WWTP dataset, it was found that bagging models perform at par or even better than ANN or SVM for the prediction and hence are suitable models that can be implemented for process control of the water treatment plants.

Index Terms—Waste water treatment plant (WWTP), ensemble models, bagging, process control.

Bharat B. Gulyani is with the Department of Chemical Engineering at BITS Pilani, Dubai Campus, Academic City, Dubai 345055, UAE (e-mail: gulyanibb@gmail.com).
Arshia Fathima is with Nanolabs, Alfaisal University, Saudi Arabia (e-mail: arshiafathima92@gmail.com).

[PDF]

Cite: Bharat B. Gulyani and Arshia Fathima, "Introducing Ensemble Methods to Predict the Performance of Waste Water Treatment Plants (WWTP)," International Journal of Environmental Science and Development vol. 8, no. 7, pp. 501-506, 2017.