Wastewater lift station monitoring and overflow alarms

At a glance

Item	Typical answer
Problem	Lift stations block, trip or lose power without warning; overflows are found by the public or by the regulator.
What is connected	Wet well level, high-level and overflow floats, pump run, fault and current, mains, rain gauge, door and gas detector, optionally a camera.
Who uses it	Wastewater control room, the on-call pump crew, the maintenance planner, the compliance officer.
Typical first scope	Five to ten stations in one catchment, including the ones with a history of spills.
Time to first screen	Typically days after the first gateway is installed; plan four to six weeks for a pilot through at least one rain event.

The situation

A sewer network relies on many small lift stations: a wet well, two or three submersible pumps, a level controller or PLC, and an emergency overflow to a watercourse. They sit in roadsides, parks and low points of housing estates, almost always unmanned. Wipes and rags wrap around impellers, grease rafts foul level sensors, and a single tripped pump during a storm is often enough to spill.

The damage is visible and expensive: pollution of a stream, flooded basements, complaints, and in many jurisdictions a reportable overflow event with its start time, duration and cause. Many stations still have only a local alarm light or an autodialler that calls one phone number. Crews visit on a schedule and clear blockages when they find them, rarely before they cause trouble.

This blueprint connects every station to Hermetry, detects the early signs of blockage and pump wear, records each overflow with its time and duration, separates storm spills from dry-weather spills, and routes alarms so the on-call crew is not buried during heavy rain.

Goals and measures

• Overflow events per station per month, with start, end and duration from the overflow float or level, and total spill hours per year.
• Dry-weather overflows: overflow events when the catchment rain gauge recorded less than an agreed amount (for example 1 mm) in the previous 24 hours. These usually mean a blockage or a pump failure, not hydraulic overload.
• Pump starts per hour = increase of the start counter per hour; above the motor's rating (often 10 to 15 for small submersibles, check the data sheet) points to a narrow level band or a failing non-return valve.
• Run time per cycle = run hours increase ÷ starts increase. A cycle that lengthens at similar inflow points to wear or partial blockage.
• Pump-down rate = wet well level fall per minute while one pump runs. Multiplied by the wet well area it gives an estimate of net pumping capacity; a falling value is an early sign of ragging.
• Station availability = time with mains present and at least one pump healthy and in auto ÷ total time.
• Mean time to respond = from alarm opened to acknowledged, and to resolved, per alarm type.

Architecture

Each station gets a licensed gateway with Hermetry Edge in the control kiosk. It reads the level controller or PLC over Modbus RTU or Modbus TCP, wires floats and contacts to its digital inputs and, where there is no PLC, reads the level transmitter on an analog input. It sends telemetry over MQTT with TLS on one outbound connection, using your own SIM and operator on a licensed 4G router or a gateway with a built-in modem. eSIM profile management from the platform, part of Hermetry Connect, is in development.

Component	Role	Notes
Licensed gateway with Edge	Reads Modbus, floats and contacts; publishes telemetry	Factory identity, claimed in one step; buffers readings while the link is down, capacity depends on the model.
Level controller or PLC	Runs duty, assist and standby pumps	Stays the only controller of the pumps.
Ultrasonic or hydrostatic level transmitter	Wet well level	4-20 mA into the PLC or the gateway.
High-level and overflow floats	Independent backup and spill detection	Wired directly to gateway inputs, not only to the PLC.
Pump protection relay or drive	Run, fault, current	Modbus where available, contacts otherwise.
Tipping-bucket rain gauge	Rainfall at a few stations per catchment	Pulse input on the gateway, counted as a total.
Fixed gas detector	Atmosphere in the wet well or kiosk	Alarm contact or 4-20 mA.
RTSP camera	Live view of the overflow or the gate	Optional; live view in Hermetry, H.264 streams.
Hermetry	Storage, alarms, rules, dashboards, SCADA	Hosted region of your choice or on premises.

Licensed devices authenticate with a certificate from the platform's own authority, and the device profile requires TLS. Where a station's PLC is already on a site Ethernet with a licensed router, the gateway reads it over Modbus TCP and the router's own web panel can be opened from Hermetry through router remote access, without a fixed IP or a VPN.

Point list

A typical three-pump station (duty, assist, standby). Register numbers are examples from a common controller map.

Point	Telemetry key	Unit	Source	Interval	Notes
Wet well level	well_level_m	m	Level controller, holding register 40001	30 s	Range 0 to well depth in the profile.
Wet well level, percent of operating band	well_level_pct	%	Controller, 40002	30 s	One alarm rule fits wells of different depths.
High-level float	hl_float	bool	Gateway DI1	on change	Independent of the transmitter.
Overflow float	overflow_active	bool	Float in the overflow chamber, DI2	on change	Start and end of each spill.
Pump 1 running	p1_run	0/1	Controller status bit or DI3	on change	Plus a heartbeat every 15 min. Send as a 0/1 number, not true/false, for history.
Pump 1 fault	p1_fault	bool	Protection relay, DI4	on change	Overload, thermal or seal leak.
Pump 1 in auto	p1_auto	bool	Controller status bit	on change
Pump 1 current	p1_current_a	A	Protection relay, Modbus	30 s	Ragging raises current at first.
Pump 1 starts	p1_starts	count	Controller counter, 40020	15 min
Pump 1 run hours	p1_hours	h	Controller counter, 40021	15 min
Pumps 2 and 3	p2_*, p3_*	as pump 1	as pump 1	as pump 1	Same keys with the pump number.
Mains present	mains_ok	bool	Phase failure relay, DI5	on change
Kiosk door	door_open	bool	Reed contact, DI6	on change
Gas alarm	gas_alarm	bool	Fixed detector alarm relay, DI7	on change	Use the detector's own relay level.
Rain total	rain_total_mm	mm	Tipping bucket, pulse input	5 min	Rain per hour and per day from the increase.
Rain, last 24 hours	rain_24h_mm	mm	Computed in a flow from the total	5 min	Read by the dry-weather rule.
Station energy	energy_kwh	kWh	Energy meter, Modbus	15 min	Optional; kWh per m³ if flow is known.
Controller communication	plc_comm_ok	bool	Gateway Modbus status	60 s
Kiosk temperature	kiosk_temp_c	°C	Gateway or small sensor	5 min	Frost and overheating of drives.

Asset model

Build the tree in Sites & Assets by catchment, because overflows and rainfall are judged per catchment:

• Catchment (free kind) with receiving_water, permit_reference and dry_weather_rain_limit_mm.
• Lift station (Site) with location, time zone, wet_well_area_m2, overflow_reference, criticality and has_generator_connection.
• Wet well (Area) and Pump 1-3 (Machine) with rated_kw, max_starts_per_hour and impeller_type.
• Rain gauge (Area) under the catchment, so rules of every station in the catchment can read it.

Place the gateway on the station asset. Because a rule may only name devices under its home, put the rain gauge under the catchment and give dry-weather rules the catchment as their home: the rule can then read both the station's overflow and the catchment's rain gauge. The tree also drives the Sites board, which colours each station by the worst status of its devices, and lets the compliance officer filter alarms and history by catchment.

Device setup

Profiles

Create one profile per station layout, for example Lift station, 2 pumps and Lift station, 3 pumps. Use the Recent message online rule with an expected interval of 30 s and declare every key with its unit and range. Put alarm rules and the command catalog on the profile, so a new station of the same layout needs no rule work at all.

Provisioning and settings

Claim each licensed gateway by scanning its code and assign the profile. Modbus addresses, serial settings, input debounce and reading intervals are managed with remote device settings: per station from the device's Settings tab, or for every station of a layout with a baseline applied by a bulk job. The job runs in waves and pauses itself if too few devices apply the change; a dry run shows the effect first. Secrets in settings are never shown; leave a secret field unchanged to keep it.

Intervals and report by exception

Floats and contacts report on change with a periodic heartbeat; level and current every 30 seconds. The gauge's running total becomes a 24-hour rain sum in the gauge profile's root flow: a Delta block turns the total into rain per reading, a Statistics block sums it over the last 24 hours, and the result is saved as rain_24h_mm. Test it in the flow test runner before deploying. A 30-second level gives a usable pump-down rate. If the gateway model supports a deadband, keep it below 1 cm for level so that pump-down rates stay accurate; if not, the profile's root flow can drop unchanged values with a report-by-exception block before they are stored.

Time stamps and outages

Overflow duration is a compliance figure, so time stamps matter. Let the gateway stamp readings only with a synchronised clock. During a link outage, licensed gateways keep readings and send them afterwards with their original times; Hermetry stores them and evaluates the overflow alarm as a backfill. A spill that began and ended while the station was offline appears in the alarm history with its real start and end, marked Backfilled, and nobody is paged for the past. A spill still open when the station reconnects is announced once.

Cameras

Where a camera watches the overflow outfall or the gate, add it as a camera device. Hermetry connects to RTSP cameras and supported recorders, plays an H.264 stream live on the station page and starts the live view only when someone opens it. The number of simultaneous live cameras follows your plan.

Dashboards

Board	Audience	Widgets
Catchment overview	Control room	Map of stations coloured by status; site table with level, pumps running, open alarms and last seen; stat cards for stations in overflow now and stations with a pump fault; alarm table filtered to Critical and Major.
Station page	Pump crew	Gauge for wet well level; state cards for each pump; line chart of level and pump currents over 24 hours; the station drawing; live camera card where fitted; activity feed of recent alarms and commands.
Pump health	Maintenance planner	Bar chart of starts per day per pump; line chart of run time per cycle over 90 days; table of stations sorted by pump current against their normal value; heatmap of pump running by hour.
Overflow register	Compliance officer	Table of overflow alarms per station with start, end and duration from the alarm history; bar chart of spill hours per month; rain per day beside spills, downloadable to Excel.
Storm view	Control room during rain	Table of stations sorted by wet well level percent; stat card for rain_24h_mm per catchment; alarm table showing only Critical.

One board with the station as an entity parameter serves every station; click a row on the overview to open its page. A wall screen in the control room shows the catchment overview through a display link: read-only, no sign-in, revocable.

SCADA screens

Draw one process drawing per station layout. The Water pack covers water, wastewater and city networks and includes lift station, screw pump, bar screen and measuring flume symbols, beside pumps, valves and tanks from the core set; see the SCADA library.

• Layout: inlet sewer, wet well with level, two or three pumps with non-return and isolating valves, rising main, overflow weir to the outfall, the kiosk with mains and generator connection.
• States: pumps show running, stopped and fault; the wet well fills with well_level_pct and turns amber above the assist level; the overflow symbol turns red and blinks while overflow_active is true.
• Second readings: each pump prints its current; the wet well prints level in metres and percent.
• Reuse: attach the drawing to the station profile from its Views button so every station of the layout gets its own view and bindings; Match everything proposes devices for unbound symbols.
• Commands on symbols: a pump symbol can carry the basic-tier fault reset command, behind a confirmation that names the pump and the command.

Rules and alarms

Alarm	Condition	Delay and hysteresis	Severity	Notified	Escalation
Overflow active	overflow_active == true	Debounce 30 s raise, 120 s clear	Critical	On-call pump crew and control room, SMS and push	Reminder every 10 min until acknowledged
Dry-weather overflow	Rule in the catchment: when overflow_active becomes true, only if rain_24h_mm is below the catchment limit, then raise alarm and notify		Critical	Pump crew supervisor, SMS; compliance officer, e-mail	Runbook: treat as blockage or failure
High level	hl_float == true or well_level_pct > 90	Debounce 60 s; hysteresis 5 %	Major	Control room, push	Rises to Critical if well_level_pct > 97
Pump fault	p1_fault == true (one rule per pump)	Debounce 2 samples	Major	Pump crew, push and e-mail	Reminder every 30 min
Possible ragging	p1_current_a > attribute p1_current_limit_a	Debounce 300 s; hysteresis 1 A; dampening on	Warning	Maintenance planner, in-app	None
Mains failure	mains_ok == false	Debounce 20 s	Critical	On-call crew, SMS	Reminder every 10 min
Gas alarm	gas_alarm == true	Debounce 5 s	Critical	On-call crew and supervisor, SMS	Runbook: no entry without personal detector
Station offline	Inactivity rule, 10 min		Major	Control room, push	Reminder every 30 min
Door open	door_open == true	Debounce 60 s	Minor	Control room, in-app	None

Lifecycle and records. Each alarm opens after its debounce, is acknowledged by the crew member who takes it and resolves when the condition clears. The alarm timeline keeps opened, severity changes, notifications, acknowledgement, notes and resolution, which gives the overflow register its start, end and handling. Resolved alarms are kept for 365 days by default; an administrator can change that.

Storms. Heavy rain lifts many wells at once. Plan for it: High level is Major and only Overflow and Mains failure are Critical; reminders stop once someone acknowledges; the same change is notified at most once in five minutes; above 20 notifications a minute Hermetry sends a digest every five minutes with a link to the filtered list. Critical alarms reach recipients even when they turned that kind of notification off, unless an administrator disables that bypass.

Maintenance. Before a crew pumps a well down by hand or desilts it, mute the station for the duration. Alarms are still recorded and marked muted, so the overflow register stays complete, but nobody is paged. Mutes end by themselves and are audited.

Automation and integrations

• Commands: the profile's catalog can offer pump.reset_fault (basic) and station.pump_down (critical, high impact, with a typed confirmation), if the gateway model and the controller support them. Each command is followed from queued to successful, failed or timed out; a timeout means no answer arrived, not that nothing happened. High-impact commands are written to the audit log.
• Manual and Auto: when a crew works on a station, switching its place to Manual in the Rules card stops every rule of that place and prevents any rule from sending commands to its devices; notify-only rules marked Also runs in Manual keep warning the control room.
• Work orders: a webhook on alarm.created sends Pump fault and Possible ragging alarms to the CMMS, which creates the job; alarm.cleared closes the loop.
• Overflow notification to third parties: where a permit requires informing an authority within hours, a flow on the alarm trigger can call the authority's or your own reporting endpoint over HTTP; most teams keep a person in that loop.
• Records: overflow events and their durations export from the alarm table and history tables to Excel. Scheduled, e-mailed reports are not built in today.

Roles and access

Who	Access	Can
Control room operator	User with the IoT operator group	See all stations, acknowledge alarms, send basic commands, switch a station to Manual.
On-call pump crew	User, scoped to their catchments by policy	See their stations and cameras, acknowledge alarms, reset faults.
Maintenance planner	User with the IoT viewer group	Read every board, export history.
Compliance officer	User with the IoT viewer group	Read the overflow register and alarm history, export.
Telemetry engineer	Admin	Profiles, alarm rules, settings jobs, drawings, mutes, command catalog.

Every acknowledgement, mute, command and setting change is recorded with who and when; the audit log is a hash chain that can be verified and exported as CSV or NDJSON. If a contractor operates some catchments, give them their own organisation or a scoped membership instead of shared accounts.

Sizing and data

As an order of magnitude for 80 stations:

• Analog values (level, currents) every 30 s in one message: 80 × 2,880 = 230,400 messages a day.
• Contacts, counters and heartbeats: about 300 a station a day, 24,000 messages a day; storms add more.
• Total around 255,000 messages a day, roughly 7.7 million a month.
• Values: about 6 analog keys × 2,880 + the rest ≈ 20,000 per station a day, 1.6 million a day for the fleet.
• Per device this is far below the default rate limits of 10 messages a second and 300 a minute.

If the message allowance is tight, a 60-second level interval halves the analog traffic while keeping pump-down rates usable. Overflow and alarm history is small; raw level and current data are the bulk. Raw readings are kept for at most 90 days, so do blockage analysis within that window (export what you need to keep) and read longer trends from hourly and daily aggregates. Treat these figures as planning estimates.

Rollout plan

Phase	Weeks	Work	Exit criteria
0. Survey	1-2	Controller maps, float wiring, kiosk power and signal per station; pick pilot stations with known spill history.	Point list per layout; overflow detection method agreed per station.
1. Pilot	3-8	Five to ten stations, one catchment, rain gauge, profiles, alarm rules, drawings.	At least one rain event recorded; every overflow cross-checked with site evidence.
2. Tuning	9-10	Adjust current limits, debounce and severities from pilot data; agree on-call routing.	Fewer than an agreed number of nuisance alarms per station per week.
3. Rollout	11-24	Remaining catchments, ten to fifteen stations a week; bulk settings jobs; cameras where wanted.	All stations reporting; overflow register produced from Hermetry.
4. Commands	after 3	Command catalog per layout, tested with each controller type.	Each command tested in a safe state on site.

Commissioning checklist

1. Gateway claimed, on the right layout profile, placed on the right station.
2. Wet well level compared with a dip measurement at two levels.
3. Each float lifted by hand: hl_float and overflow_active change in Hermetry, alarms open and notify the right people.
4. Each pump run in hand: run state and current seen; fault simulated on the protection relay.
5. Rain gauge tipped by hand a known number of times; the total increases by the expected amount.
6. Gas detector alarm relay tested with the supplier's test procedure.
7. Time stamps checked against a known event; station time zone set.
8. Link unplugged for 15 minutes with a float lifted and lowered meanwhile: the offline alarm opens, and the spill appears afterwards as a backfilled alarm with the right times.
9. Drawing bound with no symbols listed in the Problems tab; camera live view opens.
10. Runbook texts filled in for Overflow, Dry-weather overflow and Gas alarm.

Pitfalls

• Overflow from the transmitter only. Grease on the sensor hides a spill; use an independent float in the overflow chamber.
• Rain gauge in the wrong place in the tree. A dry-weather rule cannot read a gauge outside its home; place gauges under the catchment.
• Everything Critical in a storm. The crew stops reading notifications; keep High level at Major and rely on reminders and digests.
• Current limits from the nameplate. Use each pump's normal running current from the first weeks of data, written as a device attribute.
• Forgetting mutes after maintenance. Set an end time on every mute.
• Counting spills from alarm count. One long spill with a chattering float can look like many; use clear debounce and dampening, and report durations.
• Commands too early. Remote pump-down without a crew on site needs a written procedure; start read-only.

What Hermetry does not replace

Pump control, level control and motor protection stay in the station controller and protection relays and must work without the link. Gas detection for confined-space entry relies on personal detectors and the permit-to-work process, not on a fixed detector shown in Hermetry. Hermetry records overflow times and durations; the regulatory return, its review and its submission remain your process.