PdM · Budget-realistic playbook
Predictive maintenance on a budget (Philippines)
Who this is for
- Maintenance and reliability engineers
- Plant managers evaluating PdM ROI
- Field technicians starting condition monitoring
- Supervisors planning operator routes
- Sensor and analyzer suppliers
- PdM service contractors
- New engineering graduates learning PdM
What's in this guide
PdM explained: condition-based, not calendar-based
Preventive maintenance (PM) replaces parts on a schedule regardless of condition. Predictive maintenance (PdM) measures actual condition and replaces only when degradation is detected. The two are not competitors; they are different stages of maturity.
| PM (Stage 2) | PdM (Stage 3) | |
|---|---|---|
| Decision rule | Calendar or hours | Measured condition |
| Data needed | None (just schedule) | Vibration, temp, oil, ultrasonic |
| Cost per asset | Low | Medium to high |
| Catches sudden failures | No | Sometimes (electrical via thermal) |
| Catches gradual failures | Sometimes (timing-dependent) | Yes (bearings, alignment, lube) |
| Right when | Plant is reactive or first-time digital | Plant has 12+ months of logbook history |
Why most PdM programs fail in Philippine plants
Three failure patterns repeat in every Philippine plant that bought PdM before they were ready:
- Sensors before discipline. The plant buys wireless vibration sensors and a SaaS dashboard, the dashboard sends 200 alerts in month 1, nobody triages them, by month 3 the alerts are silenced. The plant paid PHP 500,000 for a system nobody uses.
- No owner. The PdM data has no single accountable person. The reliability engineer thinks the maintenance supervisor reviews it; the supervisor thinks the engineer does. After 6 months of mutual assumption, nobody is reviewing.
- Treated as a tech project. The plant manager signs off on a "digital transformation initiative" with a vendor; the vendor installs and trains; the team treats it as the vendor's problem. PdM dies the day the vendor leaves.
The fix is not better sensors. It is starting at the right stage.
Stages 1 and 2 must come first
The WorkHive 4-stage path puts PdM at Stage 3 for a reason. The two earlier stages are non-negotiable prerequisites:
- Stage 1 (Paper to Digital): 6 months of clean logbook entries with fault categories. Without this, the PdM system has no historical context to compare current measurements against.
- Stage 2 (Disciplined): PM compliance above 70 percent for at least 60 days. Without this, you cannot tell if a vibration spike is a real degradation or just the bearing PM was missed.
If your plant is not at Stage 2, do not buy sensors yet. Start with the digital logbook rollout and the PM templates. The PdM investment compounds when those foundations are solid; it evaporates when they are not.
Budget tiers: what each level buys you
| Tier | Budget (PHP) | What you get | Covers |
|---|---|---|---|
| Tier 0 | 0 | Phone vibration app + operator route via WorkHive PM Scheduler | Imbalance, looseness, gross bearing faults on slow-running assets |
| Tier 1 | 25K to 80K | Handheld vibration analyzer (entry-level) | Adds spectrum analysis, envelope demodulation for early bearing defects |
| Tier 2 | 150K to 500K | Thermal camera + quarterly oil analysis subscription | Adds electrical hotspot detection, lube degradation, contamination tracking |
| Tier 3 | 500K to 2M | Ultrasonic leak detector + portable alignment laser | Adds compressed air loss, steam trap audit, alignment-driven failures |
| Tier 4 | 2M+ | Permanent wireless sensors on top critical assets + cloud dashboard | Continuous monitoring, AI anomaly detection, automatic alerts |
Most Philippine plants we benchmark plateau at Tier 2 because the ROI of Tier 3 and Tier 4 requires PM compliance and skill matrix maturity that the plant has not yet built. Stop at the tier where your discipline runs out, not where your budget runs out.
Phone-based vibration: real or theatre?
The fair answer: real for screening, theatre for diagnosis. Modern phone accelerometers sample at 100 to 400 Hz and resolve down to about 0.01 g. This is enough to detect imbalance (1x running speed), looseness (1x and harmonics), and developing bearing defects on equipment that runs below 30 Hz (1800 RPM and slower, which covers most pumps, fans, and motors in industrial plants).
What it cannot do:
- High-frequency bearing defect signatures (typically 4 to 10 kHz): needs a real analyzer
- Orbit plots, phase analysis, modal testing: needs professional kit
- Repeatable measurement without a proper mount (phone in hand vs phone in magnetic mount changes the data)
The right use of phone vibration: monthly operator route on 15 to 30 critical assets, trend the overall velocity reading, escalate to a Tier 1 handheld analyzer when a phone reading crosses a threshold. The phone is a smoke detector; the analyzer is the fire truck.
The tool this guide is about
WorkHive Predictive Maintenance starts at Tier 0 and scales with you
The Predictive Maintenance surface in WorkHive sits inside Analytics. It reads your Logbook history, your PM compliance, and (when added) sensor data from any source: phone app, handheld analyzer export, wireless sensor API. AI anomaly detection unlocks at Stage 3 once your data has 90+ days of history. Free at the worker tier; sensor integrations roll on as you mature.
Open Predictive MaintenanceNo hive yet? Join WorkHive first (free, takes 30 seconds).
12-month rollout sequence
The rollout pattern that has worked for the Philippine plants we have advised:
- Months 1 to 3 (Tier 0): pick 20 critical assets. Install a phone vibration app. Operator does a monthly route via WorkHive PM Scheduler with the route as a recurring PM. Borrow a thermal camera for one electrical panel scan per quarter. Total cost: under PHP 5,000 for the borrowed camera rental.
- Months 4 to 6 (Tier 1): buy a handheld vibration analyzer (PHP 60K range). Train one technician as the analyst. Add quarterly oil sampling on the top 5 oil-filled assets via a third-party lab (typically PHP 800 per sample). Define escalation rules: phone reading above X triggers analyzer follow-up.
- Months 7 to 9 (Tier 2): buy a thermal camera (PHP 150K to 300K depending on resolution needed). Add ultrasonic leak detection on compressed air. Expand operator route to 50 assets.
- Months 10 to 12 (consolidate): review what worked. Calculate ROI on each tier. Document the patterns the team has found (which assets degrade in which way) so the AI assistant has training data.
- Year 2 (selective Tier 4): only for the 3 to 5 highest-criticality assets where the year-1 data showed real predictive value. Wireless sensors on a steam turbine bearing or a critical compressor make sense. Wireless sensors on a non-critical conveyor motor are theatre.
Realistic ROI expectations
SMRP benchmark: mature PdM programs achieve 35 to 50 percent reduction in unplanned downtime. Most Philippine plants in their first 18 months of PdM (Tier 0 to Tier 2) achieve 15 to 25 percent. The pattern is consistent:
- Months 1 to 6: almost no measurable improvement. The program is building data, not yet acting on it.
- Months 7 to 12: 5 to 15 percent reduction in unplanned downtime as the first caught faults get prevented.
- Months 13 to 24: 15 to 25 percent reduction as routine catches become embedded in the operating habit.
- Year 3+: 30 percent and above is possible for plants that combine PdM with strong skill matrix coverage and PM discipline.
The plants that underperform are the ones that bought too much sensor too fast. The plants that overperform are the ones that built discipline first.
Frequently asked questions
What is predictive maintenance and how does it differ from preventive?
How much does predictive maintenance cost in a Philippine plant?
Can I really do vibration analysis with a phone?
Why do most predictive maintenance programs fail?
What is the right order to roll out PdM?
Will PdM eliminate breakdowns completely?
Sources
- Society for Maintenance and Reliability Professionals (SMRP), Best Practices, 5th Edition, 2017. PdM maturity model and ROI benchmarks.
- ISO 13374, Condition monitoring and diagnostics of machines: data processing, communication, and presentation. Source for condition-monitoring data architecture.
- ISO 17359, Condition monitoring and diagnostics of machines: general guidelines. Source for the PdM decision tree.
- API 670, Machinery Protection Systems. Source for vibration thresholds on rotating equipment.
- WorkHive platform positioning, "Four Gaps One Hive" with PdM at Stage 3 of the 4-stage path. workhiveph.com
- Related WorkHive guides: Digital logbook (Stage 1 foundation) · PM templates (Stage 2) · MTBF vs MTTR