Latest / Internet of Things with Fexingo: Connected Devices, Sensors, and Industrial IoT / How IoT Sensors Are Preventing Sewage Overflows in Cities
Transcript
- Lucas: So, we talk a lot on this show about IoT sensors making things smarter — farms, factories, parking lots. But there's one application that’s literally about what flows beneath our feet, and most people never think about it until something goes wrong. Luna: Are we talking sewage? Because that’s the one that makes everyone go quiet at dinner parties. Lucas: Exactly. Sewage overflows. Combined sewer overflows, or CSOs, to be precise. In the US alone, there are about 23,000 to 75,000 overflows each year — that's raw, untreated wastewater spilling into basements, streets, and rivers. The EPA estimates it costs municipalities billions annually in cleanup and health costs. Luna: And IoT sensors are supposed to fix this? How does a sensor stop a pipe from overflowing? Lucas: It doesn’t stop it mechanically. What it does is give operators real-time data so they can act before the overflow happens. Most sewer systems today are reactive — someone calls in a backup, or a river turns brown, and then crews get dispatched. By then, the damage is done. Luna: So it's about prediction, not reaction. Lucas: Right. The key sensor is a level sensor — usually ultrasonic or radar — mounted inside the sewer pipe or at a manhole. It measures the height of the wastewater continuously. That data goes back to a cloud platform, and when the level starts rising faster than normal — say, during a heavy rain — the system can send an alert. Luna: But a rising level isn't necessarily a surprise during a storm. How do they differentiate between normal flow and a pending overflow? Lucas: That's where the analytics layer comes in. You combine the real-time level data with weather forecasts and historical flow patterns. The system builds a model of what 'normal' looks like for that pipe. If the rate of change exceeds a threshold, it flags a potential blockage or capacity issue. Luna: And then what? Someone goes out with a plunger? Lucas: Not quite a plunger, but essentially. The alert can trigger a preemptive action — like opening a diversion gate to reroute flow, or dispatching a vacuum truck to clear a blockage before it causes a backup. Some cities are even installing automated valves that throttle inflow based on sensor data. Luna: That's impressive. Give me a real city that’s doing this well. Lucas: Cincinnati is a great example. Their Metropolitan Sewer District started deploying IoT level sensors in about 200 locations across the system around 2020. They partnered with a company called Innovyze — now part of Autodesk — to build a digital twin of their sewer network. Luna: A digital twin? So they can simulate overflows before they happen? Lucas: Exactly. They feed the sensor data into the twin, run scenarios, and decide whether to pre-emptively lower water levels in storage basins or adjust pump stations. In one instance, during a major storm in 2023, the system predicted an overflow at a specific interceptor sewer. They reduced flow from upstream and avoided a spill entirely. Luna: What about Atlanta? I heard they had a big consent decree with the EPA. Lucas: Yeah, Atlanta has been under a federal consent decree since the 1990s to reduce their CSOs. They’ve invested heavily in tunneling and storage, but they also deployed IoT sensors across 70 monitoring stations. The sensors track not just water level, but also pH, turbidity, and flow velocity. Luna: So they get a water quality picture too. That makes sense for river discharge. Lucas: Right. Because when they do have to discharge during extreme events, they want to know exactly what's going into the river. The IoT data helps them document compliance with permits. And it’s not just about avoiding fines — it’s about public trust. Luna: That’s a really concrete, unglamorous problem that IoT is actually solving. I like that. Lucas: And it’s surprisingly cost-effective. The sensors themselves can cost as little as $500 to $2,000 each, plus installation. Compare that to the cost of a single major overflow cleanup — the EPA pegs the average at around $500,000 per incident. One city, I think it was Louisville, estimated they avoided $10 million in cleanup costs with a $500,000 sensor investment. Luna: That’s a twenty-to-one return. Okay, but what about the challenges? Sensors in sewers — that’s a hostile environment. Wet, corrosive, debris, rats. Lucas: Huge challenge. Ultrasonic sensors can get fouled by grease or debris. Radar is more reliable but more expensive. Some cities are using non-contact radar precisely because it can penetrate steam and is less affected by buildup. Battery life is also a constraint — most run on lithium packs that need replacement every one to two years. Luna: And connectivity? Sewers are underground, often with thick concrete lids. Lucas: Yeah, a lot of them use LoRaWAN — that’s a low-power wide-area network protocol — because it can penetrate manhole covers and transmit over a mile. Some sites need cellular gateways if LoRa isn't available. The data rate is tiny — just a few bytes per transmission — so it’s perfect for periodic level readings. Luna: So the tech is there. But is the adoption scaling? Or is it still pilot city stuff? Lucas: It’s scaling, but unevenly. Larger cities with consent decrees — like Atlanta, Cincinnati, Cleveland, and Washington DC — are leading. But smaller municipalities struggle with the upfront cost and the expertise to run analytics. Some utilities are now offering sensor as a service models, where they install and maintain everything for a monthly fee. Luna: That could help. By the way, if you're enjoying this conversation and finding it useful, you know what keeps these episodes ad-free? Listener support. You can buy us a coffee at buy me a coffee dot com slash fexingo. That's the only way we keep the lights on, and we really appreciate it. Lucas: Yeah, it genuinely helps us invest time into topics like this — the ones that are important but not flashy. So thank you if you've ever chipped in. Now, back to the sewers. One more city I want to mention: South Bend, Indiana. Luna: South Bend, really? That’s where the mayor was Pete Buttigieg, right? Lucas: Same city. They started their smart sewer project back in 2010, long before it was trendy. They installed about 140 sensors across their combined sewer system and used predictive analytics to reduce overflows by 70 percent. They even won a US Water Prize for it. Luna: So they were early adopters. What was the key insight they had that others missed? Lucas: They realized that most overflows weren't caused by sheer volume of rain — they were caused by blockages and hydraulic bottlenecks that could be managed. By dynamically controlling gates and pumps based on real-time data, they could wring more capacity out of the existing pipes. Luna: That's a great lesson: you don't always need bigger pipes, you need smarter ones. Lucas: Exactly. And that's what IoT enables. The South Bend project cost about $8 million for the sensor network and analytics, but it saved them an estimated $300 million in avoided tunnel construction. That’s a 37-to-1 return. Luna: Wow. So the business case is solid — especially for cities facing EPA mandates. But what about the privacy angle? Sensors measuring sewage — is there any concern about data misuse? Lucas: It’s a fair question, but the data is pretty innocuous — water levels, flow rates, maybe some basic chemistry. It's not like you're tracking individual household usage. Some systems do monitor upstream to detect illegal dumping, but that's about chemical signatures, not personal data. Luna: Okay. So what's the next frontier? Where does this technology go from here? Lucas: I think the big push is integrating AI for predictive maintenance — not just predicting overflows, but predicting when a sensor itself will fail, or when a pipe is about to collapse. Some researchers are working on acoustic sensors that can 'hear' blockages forming. Luna: Listen to a clog? That's wild. So the sewer system becomes a kind of nervous system — listening, feeling, reporting. Lucas: Right. And the ultimate vision is a fully autonomous sewer network: one that self-optimizes flow, self-diagnoses faults, and only calls a human when something truly needs a wrench. We're probably a decade away from that, but the sensors we talked about today are the first step. Luna: It’s amazing to think that the same basic sensor tech that tracks coffee from bean to cup is also keeping raw sewage out of our rivers. Lucas: Yeah, the hardware is often the same — it's just the application and the analytics that change. And that's what I love about IoT: it's a versatile toolkit. We just have to be clever about where we point it. Luna: Well, I feel a little smarter about what's under the street. Thanks for diving in. Lucas: Pun intended. Thanks for being here. We'll be back next time with another corner of the connected world.