Thoughts From Engineers: A Pulse on the Watershed: The Evolution of Smart Water Infrastructure Networks
The speed with which smart networks are colonizing our homes and cities’ infrastructure is proof of the technology’s adaptability to the systems of our lives. We have smart thermostats, digital surveillance systems and other Internet of Things (IoT)-driven devices to detect leaky pipes or deliver status alerts on random household appliances. Smart sensors also are increasingly being used to enhance operations in large, centralized infrastructure such as combined sewer facilities. Strategically located smart sensors collect and transmit data in real-time, and actuators control the rerouting of stormwater to designated holding areas, optimizing facility use and avoiding overcapacity and combined sewer overflow (CSO) events.
But it’s one thing to implement smart technology within a centralized sewer system and quite another to leverage a smart network of sensors and remotely activated field hardware in a large and complex watershed. And yet, all indicators suggest we’re heading in this direction. Catchment-scale stormwater management remains a compelling area of research for obvious reasons. A well-planned and executed system could check the most-damaging effects of floods by intercepting flow, diverting it temporarily to wetlands where it has time to infiltrate or incrementally releasing flow to reduce the risk of flash-flooding downstream. Notwithstanding the issues associated with broad implementation of these strategies—and jurisdictional disputes could be a major one—the idea warrants further research and analysis. We have only just started to experiment with what smart systems can do, and it’s exciting to see where our efforts will take us.
Early Adoption: Sensors on the River
The progression toward intelligent watersheds started with smart rivers. Wireless flood-monitoring networks have been active on rivers in the United States and abroad for some years now. The Town of Carey, N.C., for example, has implemented a flood-alert system (bit.ly/StreamFloodMonitoring) on river corridors within town borders via a wireless monitoring network that triggers community action when specific precipitation thresholds are reached. The Iowa Flood Information System (bit.ly/IowaFloodMonitoring) collects data from more than 200 sensors in one of the more-advanced flood-monitoring networks in the United States. These projects involving successful transmission of real-time data via wireless networks along river corridors set the scene for further development and innovation in terms of smart water infrastructure.
Real-Time Control of Flow
Combined sewer systems retrofitted with smart technology have the capability to manage incoming stormwater threatening to overwhelm the system. Configurations for real-time support systems currently in operation can differ in terms of hardware, setup and other variables, but these systems typically include a centralized control hub that monitors incoming precipitation and sewer conditions in real time via data transmitted from flow meters, sensors, rain gauges and other devices.
The sewer system in South Bend, Ind., for example, consists of 150 sensors and nine automated valves that actively manage incoming stormwater flows (bit.ly/SmartSewers). The network facilitates the rerouting of flow into existing inline storage facilities, optimizing the system’s existing capacity. As a result of this cost-efficient and innovative system, South Bend has managed to reduce the number of annual CSOs by more than 80 percent (bit.ly/CSOreductions).
Retrofitting a Watershed with Smart Analytics
Advocates of smart watersheds suggest all the elements needed to make catchment-scale water systems operational already exist; active and well-functioning systems such as those described above served as the necessary “building blocks” to later implementation at the watershed scale. Bartos et al. (bit.ly/OpenStormPlatform) use a small urban watershed in Ann Arbor, Mich., (Malletts Creek) to demonstrate how adaptive control of stormwater is possible through use of valve controllers, adaptive control protocols and cloud-based hydrologic models. The researchers argue it’s possible to not only monitor extreme storm events through strategically placed sensors, but to prevent the most-destructive events from occurring by remotely managing the system via control mechanisms—gate and butterfly valves—at key locations in the catchment.
Early analysis of the watershed identified that most runoff originated in its headwaters, characterized by significant impervious area; to a lesser degree, runoff originated in the watershed’s mid-reaches. These conditions led to the construction of several flood-control basins in the upper reaches of the catchment by local water managers.
In this study, the basins were retrofitted with remotely controlled valves. Sensor nodes attached to the valves were equipped to collect data on a variety of key metrics. Roughly 20 sensor nodes were distributed throughout the watershed, programmed to monitor site conditions and system performance. A sensor in the wetland basin was key to keeping water levels below a specific threshold; the wetland basin had overflowed during extreme storm events in the past. Data from this sensor determined the optimal time to release water from the basin located upstream.
An experiment designed to gauge how well the watershed could perform during a rain event when the different sub-elements (e.g., wetlands, upland basins, etc.) were actively managed in real-time was run. A few important findings: the actively managed watershed scenario resulted in better system performance than would have been achieved without it. Stormwater was held back at the upper basin, and roughly 19 million liters of water were eliminated from the total storm discharge. Peak flows at the watershed’s outlet were reduced by half, as indicated by the downstream hydrograph. The authors also point to reduced concentrations of suspended solids in the watershed’s downstream reaches. These are only a few of the study’s many interesting findings.
I’m compelled to consider how such a system might work in my hometown of Madison, Wis., a city famous for its beautiful, but compromised, lakes located in the heart of the city. Agricultural runoff from the contributing watersheds caused water quality in the lakes to decline. Flooding in downtown Madison also has been an issue in recent years.
The fact that these problems are now chronic—even with the area’s municipalities working cooperatively to find a solution—suggests another strategy may be needed. Could it be one driven by smart analytics? Much is clear: watershed management that draws insights from smart data as it combines the best of what green and gray infrastructure has to offer represents a truly multidisciplinary approach to water-resource management. Such a strategy is of undeniable importance in the years ahead.
About Chris Maeder
Chris Maeder, P.E., M.S., CFM, is engineering director at CivilGEO Inc.; email: [email protected].
