Qualitative estimation of SBR biological nutrient removal performance for wastewater treatment

Authors

  • Joan Colomer,

    Corresponding author
    • Control Engineering and Intelligent System Group (eXiT), Department of Electrical, Electronic and Automatic Engineering (EEEA), University of Girona, Girona, Spain
    Search for more papers by this author
  • Alberto Wong,

    1. Control Engineering and Intelligent System Group (eXiT), Department of Electrical, Electronic and Automatic Engineering (EEEA), University of Girona, Girona, Spain
    Search for more papers by this author
  • Marta Coma,

    1. Laboratory of Chemical and Environmental Engineering (LEQUIA), Institute of the Environment, University of Girona, Girona, Spain
    Search for more papers by this author
  • Sebastià Puig,

    1. Laboratory of Chemical and Environmental Engineering (LEQUIA), Institute of the Environment, University of Girona, Girona, Spain
    Search for more papers by this author
  • Jesus Colprim

    1. Laboratory of Chemical and Environmental Engineering (LEQUIA), Institute of the Environment, University of Girona, Girona, Spain
    Search for more papers by this author

Correspondence to: Joan Colomer,Control Engineering and Intelligent System

Group (eXiT),Department of Electrical, Electronic and Automatic Engineering

(EEEA),University of Girona, Campus Montilivi s/n, Escola Politecnica Superior

IV.E-17071 Girona, Spain.E-mail: colomer@silver.udg.edu

Abstract

Background

The main goal of wastewater treatment is to obtain high quality effluent. This study proposes a methodology to estimate in real-time the effluent quality in a biological nutrient removal (BNR) sequencing batch reactor (SBR) process.

Results

This is achieved by: (i) detecting the batch quality; and (ii) predicting the classification of the release according to different effluent characteristics. A principal component analysis (PCA) model is built to discern normal or abnormal behavior of the batch release. An index is given to every phase of the process by means of contribution analysis, and a fault signature (FS) is created. The FS in a classification model is associated with a biological removal quality.

Conclusion

The model is applied as a soft-sensor in real-time to new batch releases to obtain a qualitative estimate of the effluent. A correct estimation for the qualitative variables, of above 95%, would provide a reliable tool to estimate BNR performances.© 2012 Society of Chemical Industry

Ancillary