SBIR Phase I: Evaluating the Feasibility of a HAA-Meter for Affordable, Onsite Monitoring of Haloacetic Acids in Drinking Water Plants with Advanced Machine Learning Chromatographic Interpretation

PROJECT SUMMARY/ABSTRACT
Research has suggested that there is an increased risk of cancer associated with long-term exposure to drinking
water containing halogenated disinfection by-products (DBPs). Haloacetic acids (HAAs) are the second most
common class of halogenated DBPs formed during water chlorination. Every year, thousands of drinking water
treatment plants (WTPs) struggle with HAAs compliance issues often allocating significant portions of the
operating budget towards electrical and chemical means to minimize formation of the DBPs. Unfortunately, most
WTP operators adjust treatment processes without having on-site, real-time HAAs concentration data to base
their decisions on, leading to excess chemical and energy usage and higher operating costs. Standard USEPA
methods are suitable for quarterly compliance monitoring HAAs, but are often too expensive, complex, and not
well-suited for continuous real-time monitoring.
This Small Business Innovation Research Phase I project specific aims will focus on evaluating the technical
feasibility of an affordable, on-site, HAA-Meter with an integrated machine learning algorithm that will allow
automated chromatographic data interpretation and analysis. The proposed technology will drastically reduce
the cost of current available technology by half, providing WTPs an affordable option for onsite HAAs monitoring
and analysis. We will achieve this by integrating concepts, techniques, and components recently derived from
our NSF and NIEHS SBIR Phase II research projects. Additionally, the machine learning model will enable
automated interpretation of the chromatograms, evaluation of data quality, and provide Total HAAs
concentrations to the WTP operator with no user intervention, thus dramatically simplifying the analysis.
The HAA-Meter, will provide WTPs the ability to respond in real-time to minimize HAAs concentrations produced
at the WTP and in the distribution system. This lowers the risk of cancer from HAAs for communities across the
United States, large and small, rural and metro. The HAA-Meter analyzer will be fully automated and designed
for operators with skill levels consistent with typical WTP personnel. Ultimately, the result of this SBIR proposal
will be to determine the technical feasibility of reducing costs and establishing a novel machine learning algorithm
to fully automate data analysis and interpretation of the HAA-Meter.

Grant Number: 1R43ES037584-01
NIH Institute/Center: NIH
Principal Investigator: Paul Brister