Water and Wastewater Pump Stations: From Repeat Seal Failures to Predictive Maintenance

Water and wastewater pumping has a reliability profile of its own. Clean-water transfer and booster pumps run thousands of hours a year near constant duty and fail slowly — bearings, seals, worn wear rings and energy creep. Wastewater and stormwater pumps run intermittently, start and stop many times a day, rag, cavitate when wet wells draw down and sit in corrosive atmospheres; they fail suddenly. Both groups are spread across remote stations where a monthly visit is the most anyone can promise. In a water treatment plant in Vietnam, an i-ALERT pilot found a raw-water pump motor at 25 mm/s vibration — far into the ISO damage zone — traced it to misalignment with a remote FFT within four days and the team brought it below 3 mm/s with a realignment. Five of nine assets in that first assessment needed immediate action. That is a typical first result, not a lucky one.

Failure modes by station type

Station	Common failure	Why	Detection and prevention
Raw-water intake	Bearing and seal wear, cavitation at low river level	NPSH margin varies with season; silt	Vibration trending (cavitation is broadband), run-out protection, seasonal speed limits
Clean-water booster / distribution	Seal failures on pumps running at low flow at night, energy waste	Pressure control holds pumps at the far left of the curve; equal-speed parallel control	Flow-based balancing and sleep mode, minimum-flow protection, specific energy (kWh/m³) tracking
Wastewater lift station	Ragging, dry running on snore, check-valve slam	Float/level logic, rags on impellers, starts per hour	Drive-based torque monitoring and deragging, dry-run detection, ramp control; ESA for submersibles
Treatment plant process pumps (sludge, chemical dosing, backwash)	Seal and bearing failure, blocked lines	Abrasive sludge, dead-heading against closed valves during sequences	Dead-head protection interlocked with sequence; correct seal flush; vibration on bearing housings
Deep-well and borehole pumps	Motor failure, dry running on drawdown	No access to the motor; water table changes	Electrical signature analysis at the switchboard; current-based dry-run detection

Step one: let the drive do the protecting

Most stations already run VFDs. The drive knows speed, current, torque and power at every moment; combined with the pump curve those four values locate the operating point, which is enough to detect dead-heading (power drops and head rises at speed), dry running (torque collapses) and run-out (power climbs past the curve). ReliaPumpIQ does this as a module in the station panel without flow or pressure transmitters, with logic that ignores transient air on start-up so the protection does not trip falsely. For stations with two to six pumps in parallel, ReliaPumpSync replaces equal-speed control with flow balancing: a worn pump is given a higher speed so it carries its share instead of drifting towards dead-head while the newest pump does all the work. Utilities see the result in two numbers — seal replacements and kWh per m³.

Step two: see the fleet without driving to it

A wireless vibration sensor on the motor drive-end and the pump bearing housing of each critical pump gives trend data every few minutes from stations nobody visits. Set alarms from each machine's baseline, not from the ISO table alone — a pump that lives at 1.5 mm/s and climbs to 3 mm/s in two weeks is a problem even though it is "in zone B". For submersible pumps and borehole motors, where no sensor can be mounted, ESine in the switchboard reads bearing wear, rotor-bar faults and voltage unbalance from the current, and also tells the operations team how much energy each pump is using. A gateway per station and a cellular link bring everything to one screen for all stations and, through PlusSmart, to engineers who read it.

Step three: read the data you already log

Utilities have the richest data of any sector — SCADA has logged flow, pressure, level and run status for years — and the least time to read it. A Pump Health Check takes those trends plus work orders and ranks every pump by MTBF, flow as a percentage of BEP, specific energy and starts per hour, so the maintenance budget goes to the ten pumps that matter. It also exposes knock-on effects: a pressure-zone change in one district two months ago pushed a booster station into low-flow seal failures. ISO 50001 energy programmes use the same analysis to show kWh/m³ before and after.

Numbers to put in the business case

• A seal change on a 55 kW wastewater pump: parts plus labour plus crane typically USD 1,500–4,000, plus bypass pumping or overflow risk.
• Energy: a 75 kW pump running 6,000 h/year at 10 % lower specific energy saves roughly 45,000 kWh — the price of a sensor programme every year.
• Avoided overflow or dry-well flood: regulatory and clean-up costs that dwarf both of the above.

Start with the station whose work orders are thickest. PlusSmart Vietnam proposes the pilot — usually one multi-pump station with control changes and five to ten sensors — and the plant sees the trend data before committing further.

Quick answers

Can protection be added without replacing the PLC or drives?

Yes. ReliaPump modules integrate with existing drives and PLCs; a 24-pump plant was upgraded without shutdown and without original documentation. Some features such as sensorless flow control depend on drive hardware, which PlusSmart checks in the assessment.

How do we monitor submersible wastewater pumps?

With electrical signature analysis at the switchboard (ElectroSine). The motor is under water, but its current and voltage are in the panel and carry the signatures of bearing wear, rotor faults, cavitation and supply unbalance.

What does a Pump Health Check need from us?

SCADA or drive trend exports for the pumps of interest, work-order history and, if available, pump curves. Handwritten logs are acceptable. PlusSmart returns a ranked list with recommended actions.