Development of a soft-sensor for real-time estimation of E. coli concentration at the inlet of wastewater disinfection

Monitoring microbial pollution indicators requires time-consuming laboratory analyses which give results after several hours or days. A big effort is then carried out by researchers in developing sensors for bacteria detection, but so far these technologies need further testing on wastewater applications. In this work, a soft-sensor for real-time monitoring of E. coli concentration is proposed. Conventional wastewater quality indicators are tested as predictors of E. coli concentration. The soft-sensor is calibrated and tested on data from the effluent of a wastewater treatment plant before the disinfection stage. Among the tested models, artificial neural networks showed the best performances over the test data (R2 = 0.80). The soft-sensor predictions were also evaluated over an historical scenario of the predictors, where it proved to be a useful support for real-time detection of E. coli concentration and thus to control the subsequent disinfection process