2ndOpinion.FYI中文
genius.wiki

#655 2019 · Louisville Metropolitan Sewer District / Tetra Tech (Csoft real-time control) · Municipal infrastructure / water utilities

Louisville didn't build bigger pipes for the rain — it stopped treating the pipes it had as full

the problem

A network's fixed physical capacity is treated as genuinely fixed, so the standard response to it being exceeded is to build more of it, when in reality idle capacity is often sitting unused elsewhere in the same network at the exact moment another part of it is overflowing

background

Like most older American cities, Louisville's sewer system combines stormwater and wastewater in the same pipe network. During heavy rain, the combined flow regularly exceeded the fixed capacity of pipes and treatment plants, forcing untreated overflow directly into the Ohio River and local streams — a problem every part of the industry treated as a capacity shortfall to be solved by digging bigger pipes and building bigger storage tanks, the default multi-hundred-million-dollar fix.

The Louisville Metropolitan Sewer District partnered with Tetra Tech to question that framing: at any given moment during a storm, some parts of the pipe network were at capacity while other parts, elsewhere in the same system, sat with spare room the fixed-infrastructure view had no way to see or use. Tetra Tech's Csoft real-time control (RTC) software instrumented the existing network with sensors and rainfall-forecast data, then used that live picture to actively reroute flow toward whatever capacity was actually idle across the system at that exact moment, rather than routing flow along fixed paths regardless of where slack existed.

what everyone would do

The standard industry response to a combined sewer network exceeding its fixed capacity during storms was the default multi-hundred-million-dollar fix, digging bigger pipes and building bigger storage tanks, treating the network's physical capacity as genuinely fixed and the shortfall as something only more concrete and pipe could solve.

what they saw

Louisville MSD and Tetra Tech saw that the network wasn't actually full everywhere at once during a storm, some parts of the pipe system sat at capacity while other parts, elsewhere in the same network, held spare room the fixed-infrastructure view had no way to see or use. Rather than treating the shortfall as a genuine capacity problem requiring physical expansion, the fix was instrumenting the existing network with sensors and rainfall-forecast data and using that live picture to actively reroute flow toward whatever capacity was actually idle at that exact moment, treating the pipe network as a pool of shared, reallocatable capacity rather than a set of fixed, independently sized channels.

the move

Csoft continuously reads sensor data and rainfall forecasts across Louisville's combined sewer network and dynamically redirects flow in real time toward underused capacity elsewhere in the system, treating the existing pipe network as a pool of shared, reallocatable capacity rather than a set of fixed, independently-sized channels — the same total infrastructure suddenly able to absorb far more of a storm than its static design specifications implied.

why it works

Continuously reading live sensor data and rainfall forecasts and dynamically redirecting flow toward underused capacity elsewhere in the system meant the same physical infrastructure could absorb far more of a storm than its static design specifications implied, since the actual constraint wasn't total network capacity, it was the fixed routing that had never let flow move toward wherever slack genuinely existed at a given moment. Because the fix worked by better utilizing infrastructure Louisville already had rather than building new infrastructure, it delivered savings disproportionate to its cost, more than $200 million in avoided construction costs according to MSD's own chief engineer, while helping the district move toward capturing or eliminating up to 98% of combined system flow, all through software and sensors layered onto an existing physical network rather than years of excavation and concrete. The approach also proved general enough to be recognized with the 2019 INFORMS Edelman Award as a pioneering system-wide application, since the underlying insight, that a network treated as fixed-capacity may actually hold unused slack elsewhere the moment any one part overflows, applies to any shared network with uneven, time-varying load.

the payoff

Real-time control has saved MSD more than $200 million in infrastructure construction costs that would otherwise have gone toward physically expanding pipes and storage, according to MSD chief engineer Angela Akridge, while helping the district move toward eliminating or capturing and treating up to 98% of combined system flow once the related Integrated Overflow Abatement Plan projects are complete.

where it breaks

The mechanism depends on the network genuinely having spare capacity somewhere at the moments another part is overflowing, rather than being uniformly at or near capacity everywhere simultaneously, since a network with no meaningful idle slack anywhere has nothing for dynamic rerouting to capture and would still require physical expansion regardless of how sophisticated the sensing and control layer became. It also depends on having reliable real-time sensor coverage and accurate forecasting across the network, since routing decisions based on stale, incomplete, or inaccurate live data could misdirect flow toward capacity that isn't actually available, creating new overflow points rather than preventing them. And the approach requires real upfront investment in sensors, control software, and the operational sophistication to run and maintain a dynamically managed system, a genuine cost that has to be weighed against the capital expansion it's meant to defer or replace, meaning it's most valuable specifically when idle capacity genuinely exists to be captured, not as a universal substitute for physical capacity when a network is truly, uniformly maxed out.

what came after

Louisville MSD and Tetra Tech won the 2019 INFORMS Franz Edelman Award for pioneering the first major system-wide application of real-time control software to sewer network optimization, and the approach is now cited in water-utility and operations-research literature as a model for other cities facing the same combined-sewer-overflow capital-expansion problem.

references

  1. [1]Edelman AwardLouisville Metropolitan Sewer District, 2024louisvillemsd.org
  2. [2]Improving water quality with operations research: Louisville Metropolitan Sewer District and Tetra Tech awarded the 2019 INFORMS Edelman AwardPR Newswire, 2019prnewswire.com

keep it

same kind of clever