2010 R&D 100 Winner
Real-time monitoring of drinking water distribution networks is problematic due to the size and complexity of the systems. But, rapid, accurate detection of contamination is vital to providing advance warning to the public and implementing remediation and recovery systems. Sandia National Laboratories, Albuquerque, N.M., in conjunction with the U.S. Environmental Protection Agency, Cincinnati, Ohio, developed the CANARY: Event Detection Software, a free, open source software solution for water utilities.
The software monitors water quality data from any off-the-shelf water quality sensors on the network for automated event detection and routine monitoring. It performs on-line, multivariate event detection from networked sensor data. Using statistical forecasting and classification algorithms, CANARY continuously analyzes time series signals for anomalous conditions. Graphical editors help users select and adjust parameters for the algorithms, create and edit pattern libraries, and generate graphics illustrating event detection results. Java algorithms can be developed; an application programming interface connects additional algorithms.
CANARY is running on several water utilities in the United States and on the national distribution system in Singapore.
Technology
Water quality monitoring software
Developers
Sandia National Laboratories
U.S. Environmental Protection Agency
Development Team
The CANARY: Event Detection Software Development Team:
John Hall, U.S. Environmental Protection Agency
David Hart, Sandia National Laboratories
William E. Hart, Sandia National Laboratories
Terranna Haxton, U.S. Environmental Protection Agency
Katherine Klise, Sandia National Laboratories
Mark W. Koch, Sandia National Laboratories
Shawn Martin, Sandia National Laboratories
Sean A. McKenna, Sandia National Laboratories
Regan Murray, U.S. Environmental Protection Agency
Katie Umberg, U.S. Environmental Protection Agency
Eric Vugrin, Sandia National Laboratories
