Inflow and Infiltration in Lift Stations: Why Detecting I&I Matters — and How AI Is Finally Solving It
Every lift station in your system is already an I&I sensor. The problem was never the data. It was the intelligence to read it.
Every municipal wastewater utility in the country is fighting the same hidden battle. Somewhere in the network — a cracked pipe, a leaky manhole cover, a deteriorated joint, or an illegal roof drain connection — clean water is finding its way into the sanitary sewer. By the time that water reaches the lift station, the pumps don’t know the difference. They move it. The treatment plant treats it. And the utility pays for it.
This problem has a name: inflow and infiltration, commonly abbreviated as I&I. It’s one of the single biggest cost drivers in modern wastewater operations, and it’s the leading cause of sanitary sewer overflows (SSOs) during wet weather events. Yet despite its impact, I&I has historically been one of the hardest problems to detect, quantify, and locate.
Until now. This article walks through what I&I actually is, why it matters more than most operators realize, how the industry has tried — and largely failed — to detect it efficiently for decades, and how a new generation of AI-integrated monitoring built directly into lift station telemetry is finally giving utilities the visibility they’ve been asking for all along.
What Is Inflow and Infiltration?
Inflow and infiltration are two related but distinct problems that allow water that shouldn’t be in the sanitary sewer to enter the system.
Infiltration refers to groundwater that seeps into sewer pipes through cracks, faulty joints, broken laterals, and deteriorated manholes. It’s typically a slow, persistent leak that rises and falls with the water table. Infiltration tends to be highest during wet seasons and after sustained rain events.
Inflow refers to surface water or stormwater that enters the sanitary sewer through direct connections — illegal roof downspouts tied into the sewer lateral, cracked or missing manhole covers, foundation drains, catch basins improperly cross-connected, and sump pumps discharging into the wrong line. Inflow is fast. It hits the system within minutes of rainfall and disappears nearly as quickly.
Together, the two phenomena are referred to as I&I, and they have a single common effect: they take a sanitary sewer designed to carry sewage and force it to also carry clean water that the utility now has to pump, convey, and treat at full cost.
Why I&I Matters More Than Most Realize
Treating one gallon of wastewater isn’t free. Depending on the utility, the all-in cost — energy, chemicals, labor, plant capacity, conveyance, and capital depreciation — can range from a few cents to well over ten cents per thousand gallons. When ten, twenty, or fifty percent of the flow arriving at the plant is clean water that infiltrated upstream, that’s not a rounding error. That’s millions of dollars a year in larger systems, vanishing into operating budgets without anyone being able to point to a single broken pipe.
But the cost issue is only the beginning. The downstream consequences of unchecked I&I are far more serious:
The result is a quiet, compounding tax on every utility that has I&I — and almost every utility has it. Higher energy bills, more frequent pump maintenance, inflated treatment costs, and regulatory exposure — all traced back to clean water in the wrong pipe.
How the Industry Has Tried to Detect I&I (And Why It Hasn’t Worked Well)
Wastewater professionals have known about I&I for as long as there have been sewers. The problem isn’t recognition — it’s detection. Specifically, the detection methods available historically have been expensive, slow, episodic, manual, or all four at once.
Every traditional I&I detection method shares the same weaknesses: slow results, high cost, episodic coverage, and localized scope. Meanwhile, the lift stations themselves have been sitting on the answer all along.
The Lift Station: A Natural I&I Sensor Nobody Was Reading
Here’s an insight that experienced wastewater operators understand intuitively but that the industry has been slow to operationalize: every lift station is already an I&I detector. It just hasn’t had the eyes to prove it.
During dry weather, flows follow a predictable diurnal pattern — low overnight, peaks at morning and evening, moderate during the day. Pump run times settle into a rhythm. Wet well draw-down rates are stable. Cycles per hour are consistent.
During and after rainfall, that rhythm breaks. Pump run times stretch. Cycles per hour climb. Wet well fill rates accelerate. Total daily flow spikes well above the dry weather baseline. If the storm is heavy enough, the high-level alarm fires.
The shape of that breaking rhythm — how fast it diverges from baseline, how high it peaks, how long it persists after the rain stops, how it correlates with rainfall — is a fingerprint of the I&I behavior in the basin upstream of that station. A station with little I&I barely flinches in a storm. A station with heavy infiltration shows a slow, persistent elevation that lingers for days. A station with major inflow connections spikes hard and fast and drops just as quickly.
The data has always been there. What’s been missing is the ability to capture it continuously, at every station, and interpret it intelligently — without an engineer and a spreadsheet looking backwards.
How AccuDose Is Changing I&I Detection with AI-Integrated Monitoring
AccuDose was built to do what legacy systems couldn’t: deliver real-time, continuous, intelligent monitoring at every lift station — not just the biggest ones — at a fraction of the cost of traditional SCADA. The AccuDose unit sits in the control panel, connects to existing pump signals and a wet well level sensor, transmits over LTE-M cellular, and streams data into a cloud dashboard 24/7.
That alone is a meaningful upgrade. But the real shift comes from what happens to the data after it arrives.
Continuous Baseline Learning
The AccuDose platform watches every station’s behavior continuously — pump run times, cycles, wet well levels, fill rates, draw-down rates, total daily flow — and learns what normal looks like for that specific station. Every station has its own personality. Normal in a small residential basin looks different from normal in a commercial corridor. Normal in January looks different from normal in July. The platform accounts for all of it.
AI-Driven Anomaly Detection
Once a baseline is established, the AI layer continuously evaluates every cycle against expected behavior. When a station starts running longer than it should, cycling more frequently, drawing down slower, or pulling higher total daily flow than its baseline predicts — the system flags it.
Critically, the system doesn’t just flag that something is off. It begins to characterize what is off. A slow, sustained elevation after a rainfall event has the signature of infiltration. A sharp, fast spike that begins minutes into a rainfall and dissipates almost as quickly has the signature of direct inflow. The platform looks at the shape of the deviation, not just the magnitude.
Rainfall Correlation Without Extra Hardware
One of the things that makes I&I analysis traditionally expensive is installing rainfall gauges throughout a service area. AccuDose handles this differently — by leveraging the cellular network and the precise location of each installed unit, the system cross-references flow behavior against publicly available rainfall and radar data automatically. No extra hardware. No additional installation.
A utility can ask: “Which of my stations had the worst I&I response to last Thursday’s storm?” — and get an answer in seconds, ranked, quantified, and visualized. The same question a consultant would have taken six weeks and $40,000 to answer is now a click away.
Basin-Level Insight from Station-Level Data
Because the platform sees every station continuously, it can compare basins against each other and against themselves over time. A basin whose I&I response is slowly worsening month over month is one with deteriorating pipe somewhere upstream — and now there’s continuous evidence of it, with timing and severity quantified, instead of a vague sense that “things seem wetter lately.”
Utilities can prioritize CCTV inspections, line repairs, and capital projects based on actual quantified I&I contribution per basin — instead of guesswork or stale flow studies.
No Additional Hardware Required
None of this requires new sensors, new infrastructure, or capital expenditure beyond the standard AccuDose monitoring system. If the station is already on AccuDose, the I&I intelligence is layered on top of the data the system was already collecting. The I&I detection program isn’t a separate project with a separate budget. It’s a capability that comes built in with modern lift station monitoring.
What This Looks Like for an Operator
It rained overnight. The supervisor opens the dashboard on a phone over coffee. The system has already flagged three lift stations whose overnight response to the storm exceeded their normal envelope.
Two of the three stations show classic infiltration signatures — slow build, lingering elevation, drainage continuing well after the rain stopped. The third shows a sharp, fast spike that matched the peak rainfall intensity almost exactly: inflow.
No consultant. No flow study. No three-month wait. The decisions that used to require an annual capital planning meeting are now being made the morning after a storm, with data, on a phone.
Why This Matters for the Future of Municipal Wastewater
The economics of wastewater are getting harder, not easier. Federal infrastructure funding is finite. EPA enforcement on SSOs is tightening. Climate change is producing more intense storm events that stress every collection system. Utility staff are retiring, and replacement labor is scarce.
I&I reduction is one of the single highest-leverage things a wastewater utility can pursue — every gallon of clean water kept out of the sanitary sewer is a gallon that doesn’t need to be pumped, conveyed, treated, or covered by capacity expansion.
The utilities that figure this out first are going to spend less, comply more easily, defer expensive capacity projects, and stop reacting to overflows after they happen. The ones that don’t are going to keep paying the I&I tax — and that tax compounds every year.
Conclusion: Detection Was Always the Bottleneck
I&I has never been a mysterious problem. Every wastewater professional understands what it is, where it comes from, and why it matters. The bottleneck has always been detection — specifically, the ability to detect I&I continuously, system-wide, in real time, without an expensive flow study that’s already stale by the time it’s delivered.
That bottleneck is now breaking open. With AI-integrated monitoring built directly into modern cellular lift station platforms like AccuDose, utilities are finally able to see their I&I behavior the way they’ve always wanted to — every station, every storm, every basin, continuously, automatically, and at a cost that fits inside an operating budget.
The era of episodic, expensive, manual I&I detection is ending. The era of continuous, intelligent, integrated I&I monitoring is beginning. And the difference for utilities — financially, operationally, and in terms of environmental compliance — is going to be enormous.
Ready to See Your I&I Picture?
If you’re a utility wondering what your I&I situation actually looks like, the answer is no longer years and dollars away. It’s a connection away. Request a quote or call 866-310-1055 to learn how AccuDose I&I analytics can work across your lift station network.