Pillar guide·Maintenance

Preventive vs predictive maintenance: how a CMMS bridges them

Preventive maintenance uses time or runtime triggers; predictive uses condition data. A modern CMMS supports both. Here is how to set the right mix for a wastewater plant.

UtilityRadar Team May 9, 2026 8 min read

Preventive maintenance triggers on time or runtime; predictive maintenance triggers on condition data. The right CMMS supports both — and the right wastewater plant uses both. The mistake is treating them as competing strategies; in practice they are different tools for different failure profiles.

Definitions

Four maintenance strategies cover almost everything a wastewater plant does.

Corrective maintenance (CM): fix it after it breaks. Sometimes called run-to-failure when applied deliberately to low-criticality assets. Cheap on labour until the failure cascades.

Preventive maintenance (PM): scheduled tasks at fixed intervals — calendar-based (every 30 days), runtime-based (every 2,000 operating hours), or cycle-based (every 5,000 starts). The goal is to intervene before a wear-out failure occurs.

Predictive maintenance (PdM): scheduled inspection or measurement of a condition parameter — vibration, temperature, oil chemistry, motor current — with intervention triggered by threshold breach rather than calendar.

Prescriptive maintenance: PdM plus an analytics layer that not only flags the condition breach but recommends a specific corrective action with expected outcome. Rare in wastewater outside the largest utilities.

PM in a wastewater context

A typical wastewater PM programme covers the assets where wear is reasonably predictable from runtime. Hourly operator rounds: visual checks on flows, levels, alarms. Weekly walkdowns: vibration-by-feel on motors, leak checks, lubrication tops. Monthly tasks: chemical-feed pump calibration, instrument cleaning, breaker function tests. Quarterly tasks: belt tensioning, coupling alignment checks, generator load tests. 6-month tasks: motor service (bearings re-greased, insulation tested), DO probe membrane replacement. Annual tasks: UV bulb replacement, sluice-gate exercising, transformer thermography, blower overhaul.

That is roughly 80–120 distinct PM types at a mid-sized treatment works. The CMMS schedules them, the technicians close them, the reports show schedule compliance.

PdM in a wastewater context

Predictive maintenance in wastewater concentrates on five techniques against a few failure modes.

Vibration monitoring on blowers and large pumps catches imbalance, misalignment, bearing-wear signatures, and cavitation. Modern wireless sensors are $200–500 each and need a one-time calibration; a route-based handheld is cheaper but needs a trained technician monthly.

Oil analysis on gearboxes and large bearings tracks wear-metal concentration (iron, copper, chromium) and contamination (water, particulates). $30–60 per sample, quarterly on critical units, catches incipient failure 3–6 months before vibration would.

Motor current signature analysis (MCSA) picks up rotor-bar issues, eccentricity, and stator faults from current-waveform spectra. Especially useful on submerged pump motors that you cannot easily reach with a vibration sensor.

Ultrasound detects bearing-lubrication issues, valve leak-by, and air leaks in compressed-air systems — weeks before vibration shows them.

Thermography on switchgear, MCCs, and motor terminations catches loose connections and unbalanced phases — an annual sweep is cheap insurance.

💡 Where PdM pays best High-runtime, hard-to-access, expensive-to-replace assets. A blower at 95% utilisation with a 12-week lead time on a replacement rotor is the textbook case — you cannot wait for it to fail.

The transition

The honest assessment: a lot of fixed-interval PM is wasted work. The classic study (Nowlan and Heap, 1978, US Navy MSG-3) found that only ~11% of failure modes had a clear age-related wear pattern that fixed-interval PM addresses well. The other 89% had random failure distributions where calendar-based PM does little — sometimes induces failure by introducing infant-mortality risk every time the asset is opened.

Two patterns drive the PM-to-PdM transition. Low-runtime assets on fixed PM intervals — the standby pump that runs 50 hours a year does not need a quarterly bearing greasing. Move it to runtime-triggered or condition-triggered. Critical-pump-set rebuilds on calendar — pulling a duty pump every 18 months "because the manual says so" risks introducing a fault. Move to vibration-triggered with a runtime hard-stop as backup.

The transition is incremental. Most plants move 5–10 PMs per year from time-based to condition-based, starting with the highest-criticality assets where the data exists.

What the CMMS needs to support

For PM and PdM to coexist cleanly, the CMMS needs four capabilities.

Runtime ingestion: pull operating hours from SCADA or local controllers automatically. Manual runtime entry breaks down within a quarter.

Threshold-driven work-order auto-creation: when vibration on B-101 exceeds 4.5 mm/s RMS, the CMMS opens a corrective work order, assigns it to the appropriate trade, and notifies the maintenance planner. No human in the loop on the trigger.

Historian and SCADA integration: read-only access to process data so reports can correlate failures with operating regime — the pump that fails after every storm event has a different failure pattern than the one that fails on calendar.

Multiple PM trigger types on the same asset: a critical pump might have a quarterly visual inspection (calendar), an annual rebuild capped at 8,000 runtime hours (whichever first), and a vibration-triggered diagnostic order (condition). The CMMS must let all three coexist without duplicating the work.

For background on what a CMMS is and where it fits, see the CMMS pillar guide.

Mix recommendation

A typical mature wastewater plant lands at roughly 60% preventive, 25% predictive, 15% corrective measured by work-order count. Hours-weighted, the corrective slice is bigger because emergency jobs take longer per occurrence. The trend across the sector is more PdM, less time-based PM, as wireless sensor cost has dropped from $2,000+ per channel a decade ago to $200–500 today.

Two warning signs your mix is wrong. PM ratio above 80%: you are probably over-PMing assets, generating make-work, and burying the technicians who should be doing the harder PdM diagnostic work. CM ratio above 30%: your PM programme is not catching failures, either because the PMs are wrong or because the assets need PdM coverage they do not have.

⚠ Common over-correction Plants reading their first PdM case study sometimes try to instrument every asset. PdM only pays where the failure mode shows up in the condition data — a clogged screen does not produce a vibration signature. Match the technique to the failure mode, not the asset value.

For the implementation discipline that makes any of this stick, see the 90-day playbook; for the financial case behind moving from CM-heavy to PM/PdM-heavy, see CMMS effectiveness and ROI.

UtilityRadar
More
Press Esc to close · Advanced search