Artificial intelligence based monitoring system for onsite septic systems failure

Abstract

An onsite sewage system (OSS) is a complex system that takes advantage of nature's biological processes to remove harmful pathogens from wastewater and reintroduces clean water back into the natural water cycle. However, the failure of an OSS can contaminate drinking water, surface water and release hazardous pathogen and chemicals creating hazardous conditions within the local environment. This research aims to provide an artificial intelligence (AI) solution using machine learning (ML) algorithms such as logistic regression (LR), random forest classifier, and K-nearest neighbors (KNN) to understand and examine the underlying factors that could cause septic failures. Septic records from 1970 to 2019 were collected from five different counties across the State of Indiana and ML algorithms were applied to predict areas of possible septic failures. The algorithms demonstrated accuracy of approximately 80% in prediction of OSS failures. Such algorithms can assist state and county health departments in alerting homeowners of impending failures. Such an approach has the ability to not only prevent failures but also in the maintenance of a safe environment for communities reliant of OSSs. This approach was implemented and tested in the Indiana State Department of Health (ISDH).