A Benchmarking Library for Making Smart Stormwater Research Accessible
Abstract
The advent of smart cities is poised to transform the management of our built environment. Specific to water, a new generation of smart and connected stormwater systems promises to reduce flooding and improve water quality by autonomously controlling gates and valves across entire watersheds. These cyber-physical systems provide alternatives to traditional but costly concrete-and-steel construction by squeezing more performance out of existing infrastructure.
While the idea of dynamically controlling distributed stormwater systems in real-time dates back to the 1970s, the concept has only recently grown and gained global traction - in large part due to the affordability of internet-connected sensors and data services, and the broader acceptance of other autonomous systems (e.g. self-driving cars). However, relative to other fields of autonomy, smart stormwater is still early in its adoption. Broadly speaking, the "smart" component of a smart stormwater system refers to the data processing and control strategies used to guide operation of its actuators (e.g. cellularly-connected hydraulic valve). Control strategies dictate the logic embedded into these actuators, which then are operated to manipulate the water moving through the collection system to satisfy a specific performance objective (e.g. reduce flooding). The development of control strategies to meet numerous performance objectives for a variety of stormwater systems is a completely open area of research. As such, it stands to benefit greatly from researchers and practitioners from many different disciplines who can propose new visions, standards, and technologies. To make the research field of smart stormwater accessible to different disciplines, we have developed a benchmarking library that includes: An easy-to-use, open-source Python library to allow others to quickly evaluate their control strategies on a set of real-world inspired benchmarks, and An accompanying set of tutorials and examples, and corresponding web portal to promote growth of a broader community and allow for easy cross-comparison of control strategy performances. Our ultimate goal with this benchmarking library is to progress the smart stormwater field by growing its community of researchers and also promoting collaboration and friendly competition.- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2019
- Bibcode:
- 2019AGUFM.H12B..03R
- Keywords:
-
- 1821 Floods;
- HYDROLOGY;
- 1834 Human impacts;
- HYDROLOGY;
- 1847 Modeling;
- HYDROLOGY;
- 1932 High-performance computing;
- INFORMATICS