Your steel plant spends $15M–$50M annually on maintenance. Part of that replaces parts with years of life left because a calendar said so. Another part covers $2M emergency bills when equipment fails between inspections. Preventive maintenance follows a schedule. Predictive maintenance follows the data. Neither works for everything — but matching the right strategy to the right equipment is worth millions every year. Here's the framework to decide.
Preventive Maintenance
Replace on fixed intervals — time, hours, or cycles — regardless of actual condition.
Calendar-drivenSimpleDay-one value
VS
Predictive Maintenance
Replace when sensor data shows actual degradation — whether at 8 months or 36.
Condition-drivenData-richHigher precision
Side-by-Side Comparison
PM
PdM
Trigger
Time / hours / cycles
Measured condition
Between-interval failures
Blind spot
Caught 20–40 days early
Premature replacement
30–40% of parts still good
Near zero
Setup complexity
Low — schedules only
Moderate — sensors + analytics
Time to value
Immediate
4–12 weeks baseline
Gradual degradation
Approximate
Precise — tracks real curve
Random failures
Neither
Neither
Maintenance-caused failures
Higher — more touchpoints
Lower — fewer interventions
Best for
Low-cost, high-volume, regulatory
High-value, critical, variable life
Where PM Falls Short
30–40%
Parts Replaced Too Early
Nearly a third of PM-replaced components still have half their life remaining. On $8M in annual parts, that's $2.4–3.2M wasted.
60–80%
Failures Aren't Time-Based
Most failures are condition-driven or random. A 12-month PM cycle can't catch a defect that appears at month 3 and fails at month 5.
10–15%
PM Itself Causes Failures
Each unnecessary intervention risks contamination, misalignment, or reassembly error. Fewer touchpoints on healthy equipment means fewer problems.
Where PdM Has Limits
Can't Predict Sudden Failures
Contactors welding shut, cables cut, pressure spikes — zero precursor signals. PM or redundancy is the right strategy for these.
Requires Real Investment
Sensors, data pipelines, analytics, trained people. For low-criticality equipment, monitoring costs more than it saves.
Accuracy Builds Over Time
10–20% false positives initially, dropping below 8% in 6 months. PM delivers day-one value; PdM's value accelerates over time.
The Strategic Answer
Use Both — Matched to Each Asset's Risk Profile
OxMaint manages PM schedules and PdM condition monitoring in one platform, with analytics that determine the best strategy for each piece of equipment.
Six Equipment Matchups — Who Wins Where
The right strategy depends on the asset. Here are six real scenarios with a clear verdict — decisions OxMaint helps teams make across their entire equipment base.
HSM Main Drive BearingPdM wins
PM: Replace every 24 mo. Bearing often has 4+ years left. One month-16 failure caused $2.4M gearbox damage.
PdM: Vibration detected defect at month 52. $12K planned swap during roll change. Catches between-interval failures 30+ days early.
Advantage: $2.4M+ over bearing lifetime
Caster Hydraulic Servo ValvesPdM wins
PM: Rebuild every 12 mo, $85K/yr. Some at 70% life. One failed at month 8 from contamination — PM too late.
PdM: Response time monitoring. Rebuild at 15% degradation. Some at 8 mo, some past 20. Contamination caught in 24 hrs.
Rebuild costs reduced 30–40%
Coke Oven Door SealsPM wins
PM: Replace every 6 mo. Predictable cost. One leak = $50K–$200K/day in fines. Conservative schedule justified.
PdM: 130+ doors to monitor. Failure-to-violation happens in days. Savings don't justify regulatory exposure.
When failure = immediate regulatory penalty, PM wins
ESP Rapper Motors (×60)PM wins
PM: 36-month cycle. $400 × 60 = $24K. Simple. Single failure doesn't affect production.
PdM: 60 sensors at $200–$500 each. Monitoring cost ≈ parts savings. Complexity not worth it.
Low-cost + high-population = PM more economical
Ladle Turret Slewing BearingPdM wins
PM: Replace every 5 yrs at $220K. Actual life: 3–8+ yrs. Risks $220K waste AND $4M+ seizure.
PdM: Current + acoustic + position. Detects degradation 30–40 days early — at year 3 or year 8. Runs to actual life.
Expensive + variable life = PdM dramatically superior
Rolling Mill Lube OilHybrid wins
PM: Change every 3,000 hrs. Wastes clean oil. Misses contamination-degraded oil.
Hybrid: 3,000 hr baseline + oil analysis at midpoints. Healthy → extend. Degraded → change now. Best of both.
Unnecessary changes reduced 25–40%
The Decision Matrix
Two questions decide your strategy: How bad is failure? How predictable is degradation? See how OxMaint automates this classification.
Predictable Degradation
Random / Sudden
High Criticality
Predictive
Maximum ROI. Every $1 in monitoring returns $10–50 in avoided failures.
Main drives · Gearboxes · Caster hydraulics · BF compressors
PM + Redundancy
Conservative intervals. Backup systems. Strategic spare inventory.
Contactors · Safety interlocks · Protection relays
Low Criticality
Hybrid
PM schedules refined by condition checks. Extend when healthy.
Aux pumps · HVAC · Secondary drives
Run to Failure
Replace when broken. Cheaper than any scheduled maintenance.
Lighting · Redundant sensors · Consumables
Recommended Strategy Mix
35% — Predictive
100–200 critical assets under continuous monitoring. 80% of downtime risk lives here.
20% — Hybrid
150–300 assets with PM schedules refined by periodic condition data.
30% — Preventive
300–500 assets on calendar schedules. Simple, reliable, cost-effective at scale.
15% — Run to Failure
2,000+ low-cost items. Stock spares, change when broken.
Unified Platform
Four Strategies. Every Asset. Continuously Optimized.
OxMaint manages predictive monitoring, preventive schedules, hybrid intervals, and run-to-failure tracking — refined by real failure data and cost outcomes.
Financial Impact
Annual value — shifting 40% of critical equipment from PM-only to optimized mix
$6.4M
Between-interval failures prevented
$3.2M
Premature replacements eliminated
$1.8M
Maintenance-induced failures reduced
$1.4M
Labor redeployed to value-adding work
Total Annual Value
$12.8M
Investment: $300K–$600K · Payback: 1–2 months
Field Perspective
Reliability Engineering Manager — 17 years, 2.8M ton integrated mill
$11M/yr
unplanned cost under PM-only
34%
PM replacements were premature
52%↓
unplanned downtime reduction
18%↓
total maintenance cost reduction
"We caught a finishing stand gearbox bearing three weeks from catastrophic failure — five months before its scheduled PM. That single catch saved $3.2M and funded the entire predictive program. Final portfolio: PdM on 120 critical assets, PM on 300 medium, run-to-failure on 2,000+. Neither strategy alone could have delivered these results."
PdM on highest-criticality equipment first — 80% of ROI concentrates there
Retain PM on low-cost, high-population, and regulatory-driven equipment
Track maintenance-induced failures — unnecessary PM may cause more damage than it prevents
The debate has a practical answer: both, matched to equipment. Predictive where failure is expensive and degradation is detectable. Preventive where schedules are simpler. Hybrid where condition data refines PM timing. Run-to-failure where replacement beats prevention. Explore OxMaint's unified maintenance platform, and learn how ML failure detection powers predictive programs, how supply chain management connects to predictive parts ordering, and how contractor management integrates with strategy execution.
Stop Choosing. Start Optimizing.
One platform for predictive, preventive, hybrid, and run-to-failure — refined by real data and cost outcomes.
Frequently Asked Questions
Should we eliminate PM when we adopt PdM?
No. PdM replaces PM only on specific high-criticality equipment. Most plants land at 30–50% PdM, 40–50% PM, and 10–20% run-to-failure.
Which equipment should transition first?
Assets with high failure consequence (>$100K/event), reasonable failure frequency, and gradual degradation detectable by vibration, current, or oil analysis. Typically 20–40 assets qualify immediately.
What cost reduction is realistic?
15–25% total maintenance cost reduction over 2–3 years. Unplanned downtime reduction on monitored assets typically reaches 40–60%.
Can condition data extend PM intervals without full PdM?
Yes — add oil samples or vibration checks midway through PM intervals. If healthy, extend 25–50%. This hybrid approach captures 40–60% of PdM value at a fraction of the cost.
What's the typical transition timeline?
PdM on 20–40 critical assets in months 1–4. Expand to 80–120 by month 12. ROI typically appears by month 3–4 from the first prevented failure.
