Demo for an Advanced Real-time Monitoring and Decision Aide (ARMADA)
An alternative, “multi-site approach” is demonstrated here using an experimental computer program called the Advanced Real-time Monitoring And Decision Aide (ARMADA). The concept is that water quality signals at a “target site” can be “filtered” to remove variability due to known (measured) causes. The causes are represented by data that describes incoming water quality from upstream sites, and hydraulics operations such as pumping and tank cycling. The filtering is performed by empirical (curve fitting) models that predict target water quality from known causes, thereby accounting for their effects. The models’ prediction errors (the differences between the target’s measured and predicted water quality values) represent variability due to unknown causes, such as a contamination event that originates between sites. A statistically large prediction error causes a notification.
The program is in a zip file. Unzip its content to any empty folder, run Setup.exe from within the folder, and then follow the installer prompts.