A University-Utility Partnership
Nov 2017 - Nov 2018
Nov 2017 - Nov 2018
Nov 2017 - Nov 2018
Nov 2017 - Nov 2018
Quality Control
Because sensor data is noisy.
Ingest public radar products, process, and incorporate into hydraulic models and recommendation engine
Control Engine
Even after QA/QC process, there are still 100s of data points to synthesize and transform into actionable information.
Without coordination, actions can lead to unforeseen consequences throughout the network.
Take this toy example with 3 upstream control points and a downstream WWTP.
As we use storage structures in the network to mitigate CSOs and reduce wet weather peaks at the outlet of the network, the structures fill up and experience stress.
If we release stored water without information from the entire system, it could actually compound wet-weather effects downstream, or cause CSOs.
Alternatively, managing downstream objectives through the release from some assets could result in upstream flooding in others.
These problems require a structured framework to release upstream volumes to meet downstream objectives.
What if we did things differently? Coordinating actions at the system level?
Initially during a rain event, holding water upstream creates capacity downstream.
As storages fill, stress increases. We seek to distribute downstream capacity among upstream storages based on their stress.
When downstream is above a desired set point, the algorithm discourages releases from upstream.
Repeating this process continuously throughout a wet-weather event will nudge downstream to its objective, while utilizing upstream storage maximally
Decision Dashboard
No action without effective communication of recommendations to operators.
Decision Dashboard shows real-time and historical recommendations alongside system data like flow and levels.
Does it work?
Does it work?