Optimization of Wastewater Effluent

Student: Ken Foo, 2016-2017

Sponsor: EmNet, LLC, South Bend, IN

According to the EPA in 2013, $91 billion is required to upgrade and maintain all of the United States’ wastewater plants, but only $36 billion will be invested. As a result, many wastewater plants will require innovative methods to account for the deficit. Currently, South Bend’s wastewater treatment plant is spending approximately $140 million for upgrades and repairs. One solution to alleviate deficient funds and use the most cutting edge technology is to implement machine learning to model different scenarios in South Bend’s wastewater treatment plant in order to minimize costs and the concentration of chemicals in the effluent. This model will be based off of a set point to optimize the objective function, which is dependent on utilities, chemicals, labor, and several other parameters. This will be accomplished by placing sensors and controls throughout the wastewater plant to dynamically record, adjust, and optimize all the parameters saving millions of dollars while delivering an average of 48 million gallons of NPDES (National Pollutant Discharge Elimination System) certified water for 101,000 residents of South Bend.