Science Inventory

Artificial Intelligence based Prediction of Legionella Risk in Drinking Water Systems

Citation:

Dye, B., V. Gomez-Alvarez, AND L. Boczek. Artificial Intelligence based Prediction of Legionella Risk in Drinking Water Systems. American Geophysical Union, New Orleans, LA, December 13 - 17, 2021.

Impact/Purpose:

This presentation will describe the development of a machine based tool model to predict Legionella occurane in building water.

Description:

State and local government officials issued shelter-in-place recommendations (“social distancing”) and recommended the closing or reduced operation of buildings to stop the global pandemic caused by novel coronavirus disease (COVID-19). During this time, unoccupied and low-occupancy buildings might have experienced extended periods of low water demand without proper water management plans (i.e., mitigation). Periods of low or no occupancy can be challenging for building systems and may increase the risk of water system failures and other hazards for occupants. Reduced consumption of water can cause stagnant water to accumulate in building water systems. Water stagnation can lead to reduced water quality including presence of the bacteria Legionella pneumophila. L. pneumophila is a Gram-negative bacterium and is the major causative agent of Legionnaires’ disease. In the United States, reported cases of Legionnaires’ disease have grown by nearly nine times since 2000. Legionella grows best in warm and stagnated water or in building water systems that do not have enough disinfectant to prevent the growth and spread of microbes. This research will investigate the practicality of water quality parameters-based signatures as a screening tool and potential predictor of critical levels of Legionella in a building. Water quality parameters are relatively inexpensive to measure compared to extensive culture laboratory testing for Legionella. We have developed a convolutional neural network which we intend to be used to predict if a building’s plumbing or area of a distribution system is at increased risk of Legionella (i.e., action risk levels).

Record Details:

Record Type: DOCUMENT ( PRESENTATION/ POSTER)
Product Published Date: 12/17/2021
Record Last Revised: 12/28/2021
OMB Category: Other
Record ID: 353731