Wear-based inspection planning for slurry pumps

Predict helps you plan each pump inspection around actual wear, using the process and maintenance data your site already has, so you open fewer healthy pumps and catch the ones wearing fast.

Talk to us

Every pump wears differently, so why give them the same strategy?

  • The maintenance schedule is set once at commissioning and rarely revisited, usually only after a pump has failed.

  • One interval can’t fit every pump, so sites run 30 to 40% more inspections than they need.

  • Parts with plenty of life left get replaced just in case, and the wear evidence is lost.

Two slurry pumps in service at a minerals processing plant

How to enable continuous
inspection improvement

  1. 1

    Collect available data

    Predict reads the process and maintenance data the site already records, plus the wear photos taken at the pump. There are no new sensors to fit.

    • Process data: throughput and run-hours, read on a schedule
    • Maintenance data: part replacements and work orders
    • Inspection data: photos and 1–10 wear ratings, submitted through the Predict inspection app
    Data in Predict
    Process data
    Throughput & run-hours
    Connected
    Maintenance data
    Replacements & work orders
    Connected
    Inspection photos
    Captured at the pump via the app
    12 new
  2. 2

    Analyse the pump’s health

    Predict analyses the health of each pump from three inputs: every wear part against its own limit, the AI assessment of the inspection photos, and the pump’s efficiency. Together these show how the pump is wearing and how well it is running.

    • Each part against its own limit, in throughput (kt) or run-hours
    • AI wear assessment and 1–10 rating from each inspection photo
    • Calculated pump efficiency
    Inspection photo of a slurry-pump cover plate liner showing abrasive wear through to bare metal
    AI assessment 3/10 Light, even wear with plenty of service life left.
    Pump health Predict
    Efficiency 94%
    Frame plate liner 6,800 / 22,000 kt
    Cover plate liner 7,540 / 7,800 kt
    Impeller 9,200 / 15,800 kt
  3. 3

    Actionable recommendations

    Predict gives one clear recommendation for each pump, based on the part nearest its limit and the date it reaches it. Each recommendation comes with the part, limit and date behind it, so an engineer can act on it directly.

    • Bring forward when a part will reach its limit before the next inspection
    • Hold when the current schedule still fits
    • Extend when no part will reach its limit before then
    Recommendations Predict
    Bring forward
    PPC245 · throat bush reaches 7,800 kt by 02/07.
    On schedule
    PPC118 · throat bush at 82% of limit. Schedule holds.
    Extend
    PPC114 · nearest part at 44%. Defer to the August shutdown.
  4. 4

    Optimise planned inspections

    Predict places every inspection on one timeline, brings forward the pumps at risk, and lines the rest up with planned shutdowns. Each completed inspection feeds back into the data for the next cycle.

    • Inspections that already line up with a planned shutdown
    • Move only the pumps that need it; defer the rest
    • Fewer inspections overall, with at-risk pumps inspected sooner
    Schedule Predict
    Scheduled inspection Proposed new date Planned shutdown
    PPC245
    PPC118
    PPC114
    Jun Jul Aug Sep Oct Nov

    PPC245 moves earlier to stay under its limit. PPC114 defers to the September shutdown.

  5. Every inspection becomes a new data point, so the next cycle is timed on better evidence than the last.

    Start the next cycle

Results from one site
across 30 pumps in one year

$1m

returned to the maintenance budget

500h

of labour time returned

20%

fewer pump failures

We extended the pumps the data said were healthy and pulled forward the two that weren’t. A year on, the budget is back and we have had no failures. It was a number I could take to my shutdown planner and defend.
Reliability Engineer South Australian base-metals concentrator

Start with a Pump Maintenance Study

  • You send a data extract. No platform to deploy, no IT project.
  • You get back a wear picture for each pump and an optimised inspection schedule.
  • Your data is handled in confidence and stays yours.

Send a quick message