The data-driven layer of the handbook. Predictive maintenance uses historical measurements, condition indicators, operating trends and failure patterns to identify AGL asset deterioration before it reaches an unacceptable condition or causes failure — moving the framework from reactive repair to forward-looking risk management.
Philosophy: prediction is not simply collecting more data — it is identifying a change in condition over time. A single reading may look acceptable in isolation; the trend across successive readings is what reveals developing risk. This division completes the progression Preventive → Corrective → Predictive, feeding trend intelligence upward into the airport-wide condition audit.
Every predictive-maintenance topic in this division follows the same underlying sequence, regardless of asset type or indicator.
The core predictive-maintenance principle, illustrated with insulation resistance — the same logic applies to CCR current stability, transformer condition, LED degradation, connector failures and fixture failures.
| Date | Asset / Circuit | Insulation Resistance |
|---|---|---|
| Jan | Circuit A | 520 MΩ |
| Apr | Circuit A | 460 MΩ |
| Jul | Circuit A | 390 MΩ |
| Oct | Circuit A | 280 MΩ |
Each individual reading may still appear usable in isolation. Read as a sequence, the same four values show a progressive, continuous decline — the signal predictive maintenance exists to catch before the circuit fails outright.
Evaluate every trended indicator for: absolute value, rate of change, seasonal influence, sudden deterioration, repeated deterioration, and comparison against historical performance.
Predictive analysis draws on evidence already generated by Divisions 01 and 02, plus dedicated test and condition records — and is only as reliable as the data feeding it.
Grouped by data type — click any topic to expand what is trended, what to look for, and the predictive objective.
Objective: identify circuits showing progressive insulation deterioration before failure occurs.
Pattern to watch: stable → gradual deterioration → accelerated deterioration → high failure risk.
Predictive indicator: an increase in minor alarms or component interventions may indicate developing CCR deterioration.
Aligns with: transformer rating, housing, sealing, cable leads, connector, pit, water, corrosion and mechanical condition already tracked as condition information.
Objective: identify deterioration or abnormal failure behaviour before widespread failure occurs.
Formula: Fixture Failure Rate = Number of Fixture Failures ÷ Number of Installed Fixtures ÷ Defined Period. Reveals localized or product-related problems.
Aligns with: HMI, commands, feedback, alarms, events, communication, PLC, network and remote status already used as ALCMS inspection areas.
Objective: identify sections of an approach-light system requiring increased maintenance attention.
Use: the resulting data to support replacement planning.
Objective: identify abnormal operating behaviour and support efficiency / replacement decisions.
Aim: identify relationships between operating / environmental conditions and failure.
Integrates: inspection, maintenance and test history into a single asset condition history — the natural upstream data source for the airport-wide condition audit's critical-asset ranking, defect-severity trends, recurring-defect trends and replacement-priority assessment.
Classify: Normal → Watch → Warning → High Risk.
Formula: Risk = Probability of Failure × Consequence of Failure. High-risk assets receive priority for investigation and intervention.
Shift: from "replace after failure" to "replace before predictable failure."
A future AGL predictive dashboard could summarize the division's indicators at a glance — status is always based on recorded airport data, never assumptions.
| Indicator | Status |
|---|---|
| Cable IR Trend | Stable |
| Earth Resistance | Stable |
| CCR Performance | Watch |
| Transformer Health | Stable |
| Connector Failures | High |
| LED Failure Rate | Increasing |
| PAPI Performance | Stable |
| Sign Failures | Watch |
| ALCMS Reliability | Stable |
| Approach Lighting | Watch |
Every predictive output in this division is produced through the same eleven-step data pipeline.
This is where the maintenance framework becomes powerful — each division answers a different question, and feeds the next.
"What maintenance should we perform routinely?"
"What must we repair or replace because a defect has occurred?"
"What does the data tell us is deteriorating, and what is likely to fail next?"
Determine overall asset risk, deterioration and replacement priority using Division 03's trend intelligence as an upstream input.
Quick answers about predictive maintenance, based on the topics in this division.
Predictive maintenance uses historical measurements, condition indicators, operating trends and failure patterns to identify asset deterioration before it reaches an unacceptable condition or causes failure. It moves the maintenance framework from reactive repair to forward-looking risk management.
Division 03 (Predictive Maintenance) contains 25 topics in four groups: Group A Electrical Trending (7 topics), Group B Equipment Performance (6 topics), Group C Reliability Data (7 topics) and Group D Predictive Outputs (5 topics).
A single reading can look acceptable in isolation while the sequence reveals developing risk. The page's example is insulation resistance on Circuit A: 520 MΩ in January, 460 in April, 390 in July and 280 in October. Each value may look usable alone, but together they show a progressive decline that predictive maintenance is meant to catch before the circuit fails.
Absolute value, rate of change, seasonal influence, sudden deterioration, repeated deterioration, and comparison against historical performance.
Measure, Trend, Compare, Detect Deterioration, Assess Risk, Predict, Act and Verify.
Inspection and preventive-maintenance records, corrective-maintenance records, electrical test and photometric reports, CCR logs, ALCMS event logs and alarm history, component-replacement and failure records, and energy, asset-register, calibration and environmental records.
Predictive analysis is only as reliable as its data. It requires correct asset ID and circuit identification, a consistent measurement method, instrument and units, valid instrument calibration, correct date and time on every reading, accurate maintenance history and failure coding, and consistent defect classification across every asset.
Every predictive output follows an eleven-step pipeline: data collection, data validation, historical comparison, trend analysis, deterioration detection, failure-risk assessment, early warning, maintenance recommendation, planned intervention, post-intervention verification, and trend updated.
Division 01 (preventive) asks what to maintain routinely, Division 02 (corrective) asks what must be repaired because a defect occurred, and Division 03 (predictive) asks what the data says is deteriorating and likely to fail next. Division 25, the airport-wide condition assessment and audit, uses Division 03's trend intelligence as an upstream input.
Predictive maintenance is more than periodic checks on a calendar.
Build the technical knowledge to read trend data, interpret degradation patterns, and schedule intervention before a defect becomes a failure.
Anticipate the fault. Trust the data. Act before failure.
Please read before relying on any content on this page or on any record downloaded from it.