Modeling Micropollutants Removal at Wastewater Treatment Plants by Data-Driven Models: The Gap Between the Promising Potential and the Reality
Mentor 1
Marcia Silva
Start Date
10-5-2022 10:00 AM
Description
Micropollutants (MP) are a broad group of chemicals which include pharmaceuticals, personal care products, flame retardant, biocides, combustion products and chlorinated compounds, byproducts of industrial processing, and others. The removal of MPs from wastewater treatment plants (WWTPs) is currently not part of operational decision making in WWTPs but should be to decrease pollution and to improve the environmental legacy that plants have. Currently, most plants use a mechanistic model where samples are slowly collected and analyzed. These represent MP’s removal through cause-and-effect relationships of physical variables, describe the fundamental process of MPs removal, and is generally done “by hand.” Data-driven approaches are slowly starting to integrate into WWTPs, using artificial intelligence to collect data quickly, abundantly, and unbiased. In our review, we examine the usage of data-driven models for managing removal of MPs at WWTPS. To collect our data, we performed an extensive literature search for paper on MP removal in WWTPs. We then compiled the papers and noted key information in a table to compare the methods of different papers. From our search, there is an extremely limited number of WWTPs using a data driven approach in WWTPs. Most papers used a mechanistic approach and analyzed pharmaceuticals being removed. From a survey our team conducted involving over eighty WWTPs, data on the type and frequency of models used in WWTPs, MP removed, and if a data-driven model is used was collected. The switch to a data driven approach has yet to be integrated into a majority of WWTPs from our research but doing so can lead to vast improvements in MP removal in WWTPS. WWTPs that use a data-driven approach increase the evaluation of operating conditions and data collection and reduce costs which allows a quick detection of inefficiencies, abnormalities, and failures.
Modeling Micropollutants Removal at Wastewater Treatment Plants by Data-Driven Models: The Gap Between the Promising Potential and the Reality
Micropollutants (MP) are a broad group of chemicals which include pharmaceuticals, personal care products, flame retardant, biocides, combustion products and chlorinated compounds, byproducts of industrial processing, and others. The removal of MPs from wastewater treatment plants (WWTPs) is currently not part of operational decision making in WWTPs but should be to decrease pollution and to improve the environmental legacy that plants have. Currently, most plants use a mechanistic model where samples are slowly collected and analyzed. These represent MP’s removal through cause-and-effect relationships of physical variables, describe the fundamental process of MPs removal, and is generally done “by hand.” Data-driven approaches are slowly starting to integrate into WWTPs, using artificial intelligence to collect data quickly, abundantly, and unbiased. In our review, we examine the usage of data-driven models for managing removal of MPs at WWTPS. To collect our data, we performed an extensive literature search for paper on MP removal in WWTPs. We then compiled the papers and noted key information in a table to compare the methods of different papers. From our search, there is an extremely limited number of WWTPs using a data driven approach in WWTPs. Most papers used a mechanistic approach and analyzed pharmaceuticals being removed. From a survey our team conducted involving over eighty WWTPs, data on the type and frequency of models used in WWTPs, MP removed, and if a data-driven model is used was collected. The switch to a data driven approach has yet to be integrated into a majority of WWTPs from our research but doing so can lead to vast improvements in MP removal in WWTPS. WWTPs that use a data-driven approach increase the evaluation of operating conditions and data collection and reduce costs which allows a quick detection of inefficiencies, abnormalities, and failures.
