Pumping station telemetry: what to measure and how to send it
Published 5 min readLeer en español
Short answer
At a pumping station, measure levels, pressure, flow, pump status and running hours, motor current or power, faults, intrusion and power supply status. Communications are usually cellular or radio, with Modbus to the PLC and MQTT or another protocol to the control centre. Control must keep working locally if the link drops.
Key points
- Telemetry lets you learn about a failure in minutes, not when a neighbour calls or a tank runs dry.
- Measure what lets you act first: levels, pressure, flow, pump status, energy and power supply.
- A few well-prioritised alarms are worth more than many that nobody responds to.
- Choose communications by coverage and data volume, and buffer data locally when the link fails.
- Roll out in stages: one pilot station, refine, then the rest.
Why does a pumping station need telemetry?
Pumping stations tend to be remote, spread out and unstaffed. Without telemetry, failures are found late: when a reservoir runs dry, when a sewage wet well overflows, or on the next inspection round.
Telemetry measures what is happening at each station and sends it to a control centre or straight to the on-call engineer’s phone. It lets you react sooner, plan site visits with real information and build a history to guide maintenance and investment decisions.
What should you measure?
Start with what lets you act. This table sums up the usual signals:
| Variable | What it’s for | Typical sensor |
|---|---|---|
| Level in wet well, tank or chamber | Pump start/stop, overflow or dry-run alarms | Hydrostatic or ultrasonic level transmitter, float switches |
| Discharge pressure | Detect bursts, blockages or loss of prime | 4–20 mA pressure transmitter |
| Flow | Volume pumped, leaks, pump performance | Electromagnetic flowmeter |
| Pump status | Running, stopped, available, faulted | Panel contacts |
| Running hours and starts | Maintenance, spotting excessive starts | Calculated in the PLC |
| Motor current and power | Blockages, wear, energy use | Power meter or drive data |
| Faults | Overload, drive fault, motor moisture probe | Contacts or drive fieldbus |
| Intrusion | Door open, unauthorised access | Magnetic door contact |
| Power supply | Mains failure, battery or UPS status | Phase monitoring relay, UPS signal |
| Water quality (optional) | Chlorine, turbidity, pH, conductivity | Dedicated analysers |
Levels, pressure, pump status and power supply already cover most incidents. Water quality and detailed energy metering can be added later.
Which alarms matter, and which are noise?
A system with hundreds of alarms ends up ignored. A practical rule: every alarm should require a specific action from someone.
Alarms that usually matter:
- Very high level (overflow risk) or very low level (pump running dry).
- Pump fault with no standby pump available.
- Mains failure or low battery.
- Prolonged loss of communications.
- Intrusion outside working hours.
- Abnormal pressure with the pump running.
What is usually noise:
- Alerts that set and clear within seconds. Add delays and hysteresis.
- Informational events (start, stop) sent as alarms.
- Duplicate alarms from one cause: if the mains fails, you don’t need ten “pump stopped” messages.
Set priorities (critical, important, informational) and decide who each one notifies and by which channel.
How does the station talk to the control centre?
The choice depends on local coverage, data volume and whether you need remote operation or just monitoring.
| Option | Pros | Cons |
|---|---|---|
| Cellular 4G | Wide coverage, enough bandwidth for remote operation | Monthly fees, reliance on the operator, patchy rural coverage |
| LTE-M / NB-IoT | Low power, good building penetration, IoT tariffs | Limited bandwidth, availability varies by operator and area |
| LoRaWAN | Very low power, long range, private network possible | Very little data, not suitable for real-time control |
| Private radio (licensed or licence-exempt) | Independent of operators, good for station networks | Needs link surveys, masts and, if licensed, paperwork |
| Fibre or Ethernet | Highest reliability and bandwidth | Only if it already reaches the site; expensive to install |
It is common to combine them: cellular as the main route plus a backup channel for critical alarms, such as SMS.
Which protocols are used?
Inside the station, the PLC or remote terminal unit (RTU) reads sensors, drives and power meters, usually over Modbus RTU or TCP.
Towards the control centre there are several options:
- MQTT: lightweight, designed for unreliable links, a good fit for cellular networks.
- OPC UA: useful when the control centre already uses it and the link has bandwidth to spare.
- DNP3 or IEC 60870-5-104: common in water and energy utilities, with timestamping and event handling designed for telecontrol.
What matters is using open, documented protocols so you aren’t tied to a single vendor. We go into more detail in Modbus, OPC UA or MQTT: which protocol to choose.
What happens when communications fail?
At some point they will. The design has to assume it:
- Local control. Level-based pump start/stop, protections and duty/standby rotation must live in the station PLC, not in the cloud or the control centre.
- Store-and-forward. The remote unit stores timestamped data and sends it when the link returns, leaving no gaps in the history.
- Loss-of-comms alarm. The control centre should raise an alarm if a station stops reporting.
How do you keep the system powered?
If the mains fails, telemetry is exactly what tells you about it. So the PLC, router and critical sensors need backup: a UPS or a battery-backed supply sized for a few hours. At stations without a mains supply, solar power with a battery can be considered for the telemetry side.
What basic cybersecurity is needed?
A connected pumping station is critical infrastructure in miniature. As a minimum:
- Nothing exposed directly to the internet: no open ports to the PLCs.
- Encrypted outbound connections (VPN or TLS) to the control centre.
- A private APN or separate network for cellular links, where possible.
- Default passwords changed and individual user accounts.
- Firmware kept up to date and access logged.
- Remote writes limited to what is essential, with confirmation.
How do you roll it out in stages?
- Inventory: which stations you have, which signals their panels already provide and what coverage each site has.
- Pilot station: install at one, ideally the most troublesome. Tune alarms, delays and screens with the on-call team.
- Standard: settle on a standard enclosure, a signal list and a configuration template.
- Rollout: extend to the rest in batches, starting with the most critical.
- Improve: with the history in hand, add pump performance analysis or predictive maintenance.
How we approach it at Kinetifab
We start with the inventory and a pilot station, with local control that works without internet and open protocols so the data stays yours. We design the control and communications side and, where needed, the telemetry modules; any power-side work is carried out and certified by a licensed electrician. You can read more under water and industrial automation, or get in touch via contact.