Making Downtime Predictable: Wear Monitoring and Condition-Based Maintenance for Slurry Pumps
November 17, 2026

On a concentrator line, slurry pump wetted parts live in the slurry and wear far faster than ordinary rotating equipment; service life runs from a few hundred hours to a few thousand, swinging with the ore. When a mill feed pump stops without warning, the mill stops with it, and the loss accrues by the hour of throughput. The emergency repair costs more than expedited freight and overtime: no maintenance window was reserved, the spares are not on the shelf, the production plan is in pieces. What wear monitoring buys is turning that surprise into a line on the schedule.

1. Three strategies, three ledgers

Maintenance comes in three strategies: reactive (run to failure and pay all at once), preventive (replace on the calendar) and condition-based (schedule around measured wear). Published sources give directional ranges for the gap. Mature predictive-maintenance programmes are put at roughly 25–30% lower maintenance cost than reactive, with 70–75% fewer breakdowns and 35–45% less downtime, in figures associated with US Department of Energy guidance. The ranges move with management discipline, but the direction holds: fixed-interval replacement saves limited money, because half of those jobs are either too early (throwing away remaining life) or too late (the failure has already started). Condition-based maintenance saves by placing each intervention where it belongs.

2. Five signals, and vibration is only one of them

Ask about condition monitoring and most people picture vibration sensors. On a slurry pump that is exactly where a programme can miss: wet-part wear is gradual and distributed, and vibration tends to alarm only when the impeller has worn unevenly enough to unbalance the rotor, or the bearings are already in trouble, long after the wetted parts should have come out. Published industry guidance is consistent on the point: falling discharge pressure at constant speed is the primary wear indicator on a centrifugal slurry pump. As clearances open, internal recirculation grows, pressure drops first, power follows, vibration comes later still.

The full list is five signals:

  • Discharge pressure at constant speed: clearances open, pressure drops first. Readings are only comparable at constant speed and constant slurry density.
  • Motor power at constant flow: internal losses climb, current rises. Density swings confound the reading, so judge it together with density.
  • Bearing-house vibration: uneven impeller wear, misalignment, bearing distress. Log weekly; the trend says more than any single reading.
  • Wetted-part thickness: gauge shells, liners and impellers with an ultrasonic thickness meter inside maintenance windows, and extrapolate remaining life from the wear rate.
  • Bearing temperature and gland flush: overheating and abnormal flush water are mechanical precursors that corroborate the wear data.
3. From data to a maintenance decision

A complete chain has four steps. Start from a baseline: the pump's works-test curves for flow, head, power and efficiency are the reference for every later judgement. Without them, "pressure is down 5%" cannot be attributed to wear or to anything else. Set thresholds from the duty: alarm limits belong to this pump on this duty, not to a generic handbook value, or the system either cries daily or stays silent. Then trend: string the thickness readings and operating data into a wear rate, extrapolate to the minimum wall, and remaining life gets a number, which means spares can be ordered early. Finally, land it in the plan: warning windows run in weeks, so the repair slots into a normal production window instead of interrupting one.

One Australian pump-reliability service has published a telling set of results: at a tier-one mining company, eight critical slurry pumps moved from a fixed 26-week inspection interval to intervals set from each pump's own wear data. Within a year, offline inspections fell by 52% and maintenance costs by about US$308,000 a year. Six of the eight intervals stretched, because those pumps wore more slowly than assumed; two came forward, because the data showed they wore faster. A fixed interval on those two pumps is precisely the blind spot of calendar-based maintenance.

Another published figure worth keeping: when variable-speed drives let the pump run below full speed, the wear rate falls with it, and wetted-part life can extend by 20–40% in duties that do not need full speed all the time. Wear monitoring plus VFD control saves on two ledgers at once, parts and downtime.

4. Three ways these programmes go wrong
  • Instrumenting everything: monitoring spend should follow criticality. The reliability literature is consistent: reserve online monitoring for the pumps whose failure stops production, often around 20% of the fleet on the critical path, and start the rest with route readings and periodic gauging. Most of the return comes from that small share.
  • Guessing the thresholds: borrowing generic alarm limits from a manual ignores site variables such as density swings and suction conditions. Thresholds should come from the baseline and the first weeks of operating data.
  • Leaving the data off the plan: the system is installed, the trends are watched, and the shutdown calendar stays exactly as it was. The value of a warning is its lead time; if the production plan never picks that lead time up, the monitoring is just a dashboard.
5. What the supplier has to bring

Wear monitoring is not a customer-side project alone. Several pieces sit on the supplier's side of the table:

  • A baseline: if the pump was performance-tested before dispatch, the works-test curves should travel with the machine as the reference for every field reading. A pump that never saw a test rig has no baseline to give.
  • Adjustable clearances: large mill pumps carry an adjustable suction-side wear ring; when the clearance opens, restore it in place and the duty comes back without a stop. That is design leaving room for monitoring.
  • Predictable spares: same-model wetted parts available for years, with stable lead times, is what turns a trend warning into an early order. A long lead time eats the warning's lead time.
  • Reading the data together: on scheduled visits, walk the thickness records and operating data with the customer, and when the wear rate moves, check whether the duty moved with it (coarser feed, higher density, both accelerate wear).
6. The EXCELLENCE PUMP part

EXCELLENCE PUMP (石泵泵业) has specialised in slurry pumps since 1998, with more than 20,000 projects delivered, and in-house casting, heat treatment, machining and assembly together with a works performance test rig, so every pump leaves with a recorded baseline. Service follows four commitments: free pump selection and scheme design, installation and commissioning supervision, long-term spare parts supply, and scheduled follow-up visits with fault response. For customers running a monitoring system, the duty data from selection converts directly into threshold suggestions, and visits include reading the trends together.

Send us your duty parameters (flow, head, solids concentration and size, running hours). Alongside the selection review, our engineers will outline a monitoring recommendation for that duty: which signals deserve online instrumentation, which are covered by route readings, and where the thresholds should start.