Smart Cities, Smart Water: Open source tools for controlling urban water systems

Sydney Australia, April 2020

What does autonomy have to do with water?

The next generation of water infrastructure will have the ability to monitor the state of the water network in real-time and autonomously control its dynamics to achieve systems scale objectives like preventing flooding and protecting environment.

Flooding is the leading cause of natural disaster deaths worldwide. Simultaneously, untold quantities of metals, bacteria, nutrients, and other pollutants are washing-off during storms into our streams and rivers. As a result, many parts of the world are dealing with chronically impaired coastlines due to algal blooms and other ecological disasters. Many of these challenges are presently addressed through classic approaches and new infrastructure construction (larger pipes, bigger basins, storage tanks, etc.).

Network of Open-Storm sensor and actuators deployed in Ann Arbor, MI, USA.

Instead of building more, there is now an unprecedented opportunity to use IoT and CPS technologies to embed water systems with intelligence. The Internet-connected water systems of the future will control themselves (gates, valves, pumps), similar to self-driving cars, to reduce flooding and improve water quality. However, there is a dire need for more CPS researchers to help enable the next generation of data and control algorithms for smart water systems. We invite you to to join this emerging field of research and will bring you up to speed during this tutorial series.

What you will learn

The goal of this tutorial is to empower participants with state-of-the-art tools for the measurement and control of urban water systems. The tutorial is split into two parts:
1. Catch up: Overview of IoT sensing technologies, CPS Python simulation frameworks, and case studies on real-world smart water systems.
2. Get stuff done: Learn the Open-Storm.org suite of open source tools to help eliminate flooding and improve water quality through the real-time control of gates, valves, and pumps.

Abstracting the physical stormwater systems as a network of interconnected elements, enables us to develop algorithms for controlling their behavior during storms to prevent floods.


Audience: Students, practitioners, and researchers seeking to apply their control algorithms and technologies to the management of real-world water systems.

Schedule

This tutorial will comprise a half-day session, followed by an optional hackathon. The tutorial will constitute a combination of talks, hands-on sessions, and an overview of real-world case studies.

  • Overview: [1hr] Initial part of the tutorial will focus on providing the participants an overview of smart water systems, test beds, open-storm hardware, IoT architectures, and algorithms for controlling urban water systems.
    • State of the art of Smart Water Systems, Prof. Branko Kerkez, University of Michigan.
    • The Open Storm Sensor Actuator Stack, Dr. Sara Rimer, Argonne National Laboratory
    • Overview of Cloud Services, Brooke Mason, University of Michigan
  • The open-storm control toolbox: [3hrs] (Abhiram Mullapudi and the Open Storm team): We will host a hands-on session on using open-storm’s Python toolbox and benchmarking framework (pystorms) for developing real-time control algorithms for urban water systems. Participants will only need their laptop and internet access. In this session, participants will learn step-by-step exercises to:
    • Control a single stormwater basin in real-time, with the goal of reducing flooding
    • Control a system of water storage assets in series and parallel
    • Control the behavior of entire, city-scale stormwater networks for mitigating flooding
  • Optional, afternoon office hours and hackathon: [4hrs] In this afternoon session we will work with interested participants to help port their control algorithms and control ideas to the Python toolbox. The event will be held as a mini-competition, with a token prize going to the group whose control algorithm minimizes flooding in a real-world inspired simulation. Our team of experienced control designers will work with the small teams to answer questions and support their control algorithm development.

Optional reading

  • Smarter Stormwater Systems: Branko Kerkez, Cyndee Gruden, Matthew Lewis, Luis Montestruque, Marcus Quigley, Brandon Wong, Alex Bedig, Ruben Kertesz, Tim Braun, Owen Cadwalader, Aaron Poresky, Carrie Pak; American Chemical Publications
  • Open storm: a complete framework for sensing and control of urban watersheds: Matthew Bartos, Brandon Wong, Branko Kerkez; Environmental Science: Water Research & Technology
  • Integrated stormwater inflow control for sewers and green structures in urban landscapes: Nadia Schou Vorndran Lund, Morten Borup, Henrik Madsen, Ole Mark, Karsten Arnbjerg-Nielsen, Peter Steen Mikkelsen; Nature Sustainability
  • Hybrid modeling and receding horizon control of sewer networks: Bernat Joseph‐Duran, Carlos Ocampo‐Martinez, Gabriela Cembrano; Water Resources Research