Local-scale Stormwater Inlet Monitoring System
Aaron Katz, Jie Li, Zoe Kerrich
Overview
LoSIMS, a local-scale stormwater infrastructure monitoring system, provides an opportunity to address street flooding with a low-cost, distributed sensor network implemented in stormwater inlets across Philadelphia.
Philadelphia has significant problems with stormwater management – and the city chose over a decade ago to emphasize improving stormwater management through expanding green stormwater infrastructure – but doing so requires that existing gray stormwater infrastructure is also functioning.
A simple, high-granular, and local-scale solution may alleviate many community flooding problems. For this, LoSIMS focuses on ensuring that stormwater inlets across the city are functioning well – clear of trash and other potential obstructions. Keeping stormwater inlets clear can improve stormwater management and reduce urban flash floods.
Our sensor system will detect when water is not flowing into stormwater inlets properly, and do two things: (1) notify the city and local residents that the inlet requires clearing and; (2) during high rain events, inform local residents of increased risk of localized flash flooding.
An ancillary benefit of this system is that it also has the potential to provide long-term data on where urban flooding events happen most frequently in the city, with high spatial specificity.
Design Concept
Our pilot design deals with a particular type of stormwater inlet on the curb as shown below. The product consists of the following parts:
- Smart cover, which integrates the central processor, cellular transmitter, battery pack, and four ultrasonic distance sensors. The four distance sensors map out the surface underneath in the basin.
- LED at the edge of the smart cover that lights up when the drain is blocked.
- Two sets of flow meters combined with turbine generators, one outside the inlet and one inside the pipeline. The flow meters measure flow rates running through them, while the turbine generators generate electricity for the battery pack, powering the entire system.
- Connecting wires protected in waterproof tubes.
How do these sensors work together? Below diagrams the working pattern of LoSIMS.
The system monitors two outcomes: storm accumulation and drain blockage.
- Storm accumulation is detected through the two flow meters in and outside the inlet. If everything runs smoothly during a rain, the flow rate of street runoff should match with that down the drain’s pipeline. These two flow rates are calibrated beforehand to figure out what the “match” is. Then, if the two flow rates do not match, it indicates that storm is accumulating in this section of the street.
- Drain blockage. Storm accumulation occurs either because the inlet is at capacity or because of drain blockage. The four ultrasonic sensors are the “surface mappers” that can detect the nature of the substance underneath. If, for example, the sensors detect a ragged surface below, this suggests that trash are blocking the basin. Or, if the water surface is too high, it directly shows that the inlet is at capacity.
Once either outcome is detected, the system sends signals to residents, the community, and the City via the following channels:
- Cellular transmission to cloud dashboard. The signal is sent through the cellular transmitter on top of the Smart Cover to an online dashboard that both the community and the City’s water department has access to.
- In-situ LED. The LED strip lights up when the drain is blocked. This is both to signify field workers which drain goes wrong and to notify residents to be prepared or to report.
Prototype Details
The prototype model consists of a smart drain cover, which integrates the processor, the cellular transmitter, the battery, four ultrasonic distance sensors, and an LED strip, two flow meters, one outside the inlet on the street and another down in the pipeline, and necessary wire connections.
The system detects two outcomes: storm accumulation and drain blockage. Flow accumulation is detected by comparing the readings from two flow meters, one detecting water flow on the street, and another detecting water flow down the drain. If there is a mismatch between the two readings, it means that water is not properly flowing down the drain, and that storm may be accumulating in the community. Then a warning of precaution is sent to the community and individuals.
Storm accumulation may be caused by drain blockage. To measure this we have multiple ultrasonic distance sensors to map out the surface inside the drain inlet. It can, for example, detect if the water level is too high. If it detects a ragged surface and if the water flow down the pipe is too weak, then this drain may be blocked by trash. In this case, a signal is sent to the water department and the community, who will send out personnel to deal with the situation.
Pilot Project Siting
Picking a site for piloting this project is an important step in implementation – it provides an opportunity to work through real-world problems that come up in the process. Based on information from Philadelphia’s draft 2022 Hazard Mitigation Plan, we’ve decided to focus on a neighborhood in North Philadelphia near Tacony Creek
There are areas in this neighborhood with relatively high flood risk – our hope is that in working with a specific neighborhood or even city block, we can make sure we’re installing our pilot in the best location for that neighborhood, and it provides an opportunity to refine our approach to community interface with the project.
City/Community Interface
The signals of drain blockage or flood accumulation are sent to the Cloud dashboard. The above shows a mockup dashboard app used by the City’s water department. Through this dashboard, the person in charge can inspect sensor readings coming from each vulnerable inlet around the city that have LoSIMS installed. The app also shows whether this inlet requires fixing and facilitates the office employee to communicate with on-site workers regarding the repair. Data is also stored for long-term analysis use.
On the other hand, notifications are also pushed to the phones of residents that subscribe to this service. In this way, residents can be prepared when their section of the street is expected to get flooded; for example, they can arrange to arrive home early or quickly go grocery shopping. Residents are also encouraged to report the situation to the community and the City to put more pressure to the repair process.
Design Considerations
Further incorporating stakeholder input, constraints, and needs will be critical to ensuring a successful and impactful storm water monitoring system.
- Municipal oversight: What are the realistic operational constraints for active municipal management and maintenance?
- Community: What might engagement/education with local community groups look like? How will residents interact/respond to a smart-drain system on their block?
Next Steps – Implementation
In scaling out this physical prototype, there are several more improvements to get this project up and running in the streets.
- Robust design. There are many things to consider to transform the fragile prototype into a robust real project. They include:
- Power supply: the system requires a tested electricity supply system powered by micro-hydroelectric turbines. These turbines only work when there are flows, but the system could stay dormant most of the time to save power, and is only triggered when rain starts.
- Waterproofing: the system will need robust waterproof sensors and reasonable housing to withstand flood levels.
- Installation: the system will benefit from even further simplified sensor packages within modified manhole cover.
- Connectivity: while the prototype uses WiFi, we envision the real product to use LoRa or cellular communication, a more power-saving and robust system.
- Community scaling. To determine pilot locations, we need to integrate community feedback with historical flood data/analysis. The neighborhoods should be actively involved and educated on drain maintenance and blockage prevention.
- Data analysis. Prior data analysis is needed to determine thresholds by which to identify an event of flood accumulation or drain blockage worth alerting. We may also consider sensor redundancy – incorporating local rain gauge to improve system accuracy.