An integrated machine learning framework for effluent quality prediction in Sewage Treatment Units

ABSTRACT

The influent and effluent data from wastewater treatment plants being highly correlated with multi-variable coupling and time-varying features may degrade the performance of conventional soft sensors over time. Adaptive strategies based on just-in-time learning (JIT), moving windows (MW), and time difference (TD) are used in this work to develop an adaptive soft sensor. Multi-output Gaussian-process regression (MGPR) is selected and hybrid methods such as TD JIT, MW TD, and JIT MW TD along with TD and MGPR methods are implemented. Data from the benchmark simulation model No.1, closed-loop architecture after applying PI controller, and real-time data from the Rithala Plant of Delhi are obtained. The improved error percentage is 15.03% for total phosphorus (open-loop) using the JIT TD method when compared with the MW TD method. Fair results are observed with JIT TD on real time data with a strong correlation between predicted and observed values, above 0.8 for any variable being estimated.

KEYWORDS:

• Activated sludge models
• benchmark simulation models
• machine learning
• multi-output gaussian process regression
• sewage treatment

Disclosure statement

No potential conflict of interest was reported by the authors.

Authors’ contributions

Chandra Sainadh Srungavarapu conceptualized the main idea and carried out the simulation studies based on the benchmark simulation data and contributed to write the manuscript. Abdul’s contribution was carrying out simulation studies and analyzing the results. E. S. S. Tejaswini conceptualized the model and performed simulation studies. Sheik Mohammed Yousuf conceptualized the model and performed simulation studies and proof reading of the manuscript. Seshagiri conceptualized the core idea and methodology, supervised the other authors for writing of the simulation of codes and also proof reading of the manuscript. All authors read and approved the final manuscript.

Data availability statement

The authors agree that sharing the codes will help other researchers in the field. . The Google Colab codes used in this paper are available upon request. Please contact Chandra Sainadh Srungavarapu ([email protected], [email protected]) at National Institute of Technology, Warangal, India, Prof. Seshagiri Rao Ambati ([email protected]) at the Indian Institute of Petroleum and Energy, Visakhapatnam, India.

Availability of data and materials

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Highlights

Adaptive strategies (machine learning) such as JIT, TD, TD MW, JIT TD MW and TD JIT are implemented on WWTP data based on influent and effluent data.

• Both open-loop and closed-loop frameworks are considered while collecting the influent and effluent data.

• On comparison, TD JIT is observed to show optimal solution with less than 0.05 error % for predicting any variable in the test dataset. Other methods also produced good quality predictions.

• The methods are implemented on a real-time data from WWTP in Delhi, India and the methods fairly predicted with more than 90% accuracy.

• The improved error percentages are of 15.03%, for total phosphorus (open-loop) using JIT TD method when compared with MW TD method.

Supplementary data

Supplemental data for this article can be accessed https://doi.org/10.1080/1573062X.2023.2183137

Additional information

Funding

This research is funded by a grant from Department of Science and Technology (DST), Government of India under Water Technology Initiative (WTI) Scheme with grant no: DST/TMD/EWO/WTI/2K19/EWFH/2019/189.