Stopping sewer overflows: monitor the collection network, not just the works

Who this is for

Municipal sanitation and water-services engineers and operations managers responsible for a network of sewage pump stations, rising mains and gravity sewers feeding a treatment works — the people who get the after-hours call when a station overflows and sewage reaches a river.

The works gets the attention. The pipes that feed it overflow.

Most monitoring effort, and most regulatory scrutiny, lands on the wastewater treatment works — final effluent quality, the discharge point, the licence. But the works is the end of the line. The bulk of raw-sewage spills happen in the collection system before the flow ever reaches the inlet: at an unmanned pump station, a surcharged manhole, a burst rising main. You can have a compliant works and a network that is spilling daily. Monitoring only the works is monitoring the one place the sewage is already under control.

Why pump stations fail silently

A sewage pump station is usually unmanned, often remote, and its failure modes are invisible from the outside until the overflow is already running down the road. The common ones:

• Tripped or blocked pumps. A pump trips on overload, a rag or wipe blocks the impeller, or the duty pump fails and the standby never starts. Sewage keeps arriving; the wet well fills.
• Full wet wells. A station built for yesterday's load, or one with a degraded pump, simply can't keep up. The level climbs to the overflow weir and spills — and the only sign was the level, which nobody was reading.
• Rising-main blockages and bursts. A blockage, air-lock or burst on the pumped line backs the station up or dumps sewage along the route of the main. Pump run-hours climb while nothing actually moves.
• Float-switch failures. The float that controls or alarms the pump fouls or sticks in sewage, so the one local safeguard quietly stops working — often discovered only at the next spill.

One stalled station can spill thousands of litres an hour of raw sewage. The interval between failure and discovery — typically the next scheduled site visit — is the entire problem.

Load shedding stops the pumps. The sewage doesn't stop.

This is the South African failure mode that turns a manageable network into a recurring pollution source. When the grid drops, the pumps stop — but flow into the wet well carries on at full rate. A two- to four-hour stage of load shedding on a station with limited wet-well buffer becomes a spill on a fixed schedule. Without a power-loss alert and a live wet-well level, operations doesn't know which of dozens of dark stations is closest to overflowing, so a generator or vacuum tanker goes to the wrong one. Knowing which stations lost power, and how full each is, is the difference between a planned response and a clean-up.

Vandalism and cable theft

Pump stations are soft targets: remote, unmanned and full of copper. Cable theft strips the supply, motor windings and control wiring; panels are vandalised or stripped for scrap. The station goes dead and stays dead until the next visit — and now the failure isn't a tripped breaker you can reset remotely, it's a station that can't pump at all. A station that suddenly goes dark is often the first sign of theft, which is why a power-loss alert doubles as a security alert, and why sensing that runs independently of the main panel keeps reporting even after the panel is gone.

Inflow & infiltration: the wet-weather overload

Ageing, cracked sewers and illegal stormwater connections let rain and groundwater pour into the collection system. This inflow & infiltration (I&I) can multiply dry-weather flow several times over in a storm, overwhelming pump stations and the works at exactly the same moment. The result is wet-weather spills across the network and a works that floods every time it rains. You can't rehabilitate what you can't localise — and the signature of I&I is in the data: sections whose flow and level spike abnormally with rainfall. Comparing dry-weather and wet-weather patterns across the network points rehabilitation at the worst offenders instead of the whole catchment.

What to monitor

The collection network needs different instrumentation from the works. The core set:

• Wet-well level — continuous level with high-level and rate-of-rise alarms. A fast rise with the pumps supposedly running is the earliest sign a pump isn't keeping up, well before the overflow weir.
• Pump run / fail / trip & power status — duty/standby state, run-hours and power presence, so a stalled pump or a lost feed registers the moment it happens, not on the next visit.
• Overflow / spill detection — level sensing at overflow weirs, repeat-offender manholes and outfalls that flags sewage escaping the network — the alert that lets you contain a spill in minutes.
• Rising-main pressure — pressure on the pumped line reveals a blockage, air-lock or burst before it backs the station up.
• Inflow & infiltration — dry- versus wet-weather flow and level patterns that localise where stormwater and groundwater leak in.
• Power & security — load-shedding and power-loss alerts plus door and panel intrusion sensing for stations exposed to outages and theft.

Why this is a Green Drop and DWS pollution issue

A sanitary sewer overflow isn't just an operational nuisance — it's a public-health and environmental incident. Raw sewage reaching a watercourse is a pollution event that can fall under Section 19 of the National Water Act (the duty to prevent and remediate pollution), and the state of the collection system is assessed under the Green Drop programme. Repeated, undetected overflows from the network are exactly the kind of finding that costs a municipality its Green Drop standing and exposes it to directives and penalties. Catching a spill as it starts turns a reportable pollution event into a contained call-out. Regulatory references here are for orientation; confirm current obligations against your own licence and the latest DWS guidance.

How to put it in place

The sequence we follow on a collection-network project:

• 1. Rank by consequence. Instrument the stations whose overflow reaches a river, a school or a main road first — most networks have a handful of repeat offenders where monitoring pays for itself fastest.
• 2. Read what's already there. Tie into existing float switches, panel signals and flow meters; add sensing only where there's a gap, rather than replacing the control gear.
• 3. Add spill detection. Level at weirs, problem manholes and outfalls so an escaping flow is caught at the boundary of the network.
• 4. Alarm to the responder. High level, rate-of-rise, pump failure and power loss to WhatsApp or email, so the standby team rolls with the right fix — pump, tanker or generator.
• 5. Back it with solar and edge buffering. Log locally and run on solar so the record and the alarms survive load shedding — the exact window when spills happen.

Why solar and edge buffering aren't optional here

The spill most often happens during the outage. If the monitoring depends on the same grid feed as the pumps, it goes dark precisely when you most need to know the wet-well level is climbing. Solar-backed sensing with local edge buffering keeps level, alarms and the record alive through load shedding and through a stripped panel, so the network stays visible when the station itself is down. We treat it as a requirement, not a nice-to-have. The same approach underpins our solar-powered remote monitoring on off-grid sites.

This is the backbone of our sewer network & pump station monitoring, feeding straight into treatment-works monitoring so the works anticipates its incoming load. It runs on the addaNet platform, so sanitation sits alongside the rest of the operation, and it's a core reason municipalities and water utilities work with us.

Frequently asked questions