Clean Technology (청정기술)

The Korean Society of Clean Technology (한국청정기술학회)
권호 표지
DOI QR코드

DOI QR Code

Stochastics and Artificial Intelligence-based Analytics of Wastewater Plant Operation

  • Sung-Hyun Kwon (Department of Marine Environmental Engineering, Engineering Research Institute Gyeongsang National University) ;
  • Daechul Cho (Department of Energy&Environmental Engineering, Soonchunhyang University)
  • Received : 2023.05.15
  • Accepted : 2023.06.06
  • Published : 2023.06.30
Download PDF
⟨ Previous Next ⟩

Abstract

Tele-metering systems have been useful tools for managing domestic wastewater treatment plants (WWTP) over the last decade. They mostly generate water quality data for discharged water to ensure that it complies with mandatory regulations and they may be able to produce every operation parameter and additional measurements in the near future. A sub-big data group, comprised of about 150,000 data points from four domestic WWTPs, was ready to be classified and also analyzed to optimize the WWTP process. We used the Statistical Product and Service Solutions (SPSS) 25 package in order to statistically treat the data with linear regression and correlation analysis. The major independent variables for analysis were water temperature, sludge recycle rate, electricity used, and water quality of the influent while the dependent variables representing the water quality of the effluent included the total nitrogen, which is the most emphasized index for discharged flow in plants. The water temperature and consumed electricity showed a strong correlation with the total nitrogen but the other indices' mutual correlations with other variables were found to be fuzzy due to the large errors involved. In addition, a multilayer perceptron analysis method was applied to TMS data along with root mean square error (RMSE) analysis. This study showed that the RMSE in the SS, T-N, and TOC predictions were in the range of 10% to 20%.

Keywords

Acknowledgement

This paper was supported by the Soonchunhyang University. Authors thank for it.

References

  1. SOOSIRO, 2022, http://www.soosiro.or.kr
  2. Koprea Environment Corporation, 2022, http://www.keco.or.kr
  3. Woongburi, P. and Park, J. K., "Big Data Analytics from a Wastewater Treatment Plant," Sustainability, 13, 12383 (2021).
  4. Jawad, J., Hawari, A. H., and Zaidi, S. J., "Artificial Neural Network Modeling of Wastewater Treatment and Desalination using Membrane Process: A Review," Chem. Eng. J., 419, 129-540 (2021).
  5. Im, J. M. and Joo, H. K., "Big Data Analysis and Forecasting of Sewage Treatment Facilities using Machine Learning," Kor. Inst. Ind. Eng., 22, 1316-1322 (2021).
  6. Jeong, H. S., Lee, S. H., Shin, H. S., and Song, E. Y., "Analysis and Prediction of Sewage Components of Urban Wastewater Treatment Plant using Neural Network," J. Kor. Soc. Environ. Eng., 28, 3038-315 (2006).
  7. Seiffert, U., "Multiple Layer Perceptron Training using Genetic Algorithms," Eur. Symp. Art. Neur. Network, 159-164 (2021).
  8. Lee, S. M., Park, K. D., and Kim, I. K., "Comparison of Machine Learning Algorithms for chi-a prediction in the middle of Nakdong River," J. Kor. Soc. Water Wastewater, 34, 277-288 (2022). https://doi.org/10.11001/jksww.2020.34.4.277