The analytical and statistical intelligence layer of the AGL Failure Database. Divisions 01–20 classify the failure, Division 21 establishes root cause, and Division 22 determines severity and criticality. Division 23 converts all of those records into reliability statistics, trends, comparisons, hotspots and management decisions — built around measurable, repeatable data rather than subjective assessment.
Core logic — Failure Data → Reliability Metrics → Trend Analysis → Hotspot Identification → Reliability Decision → Corrective/Preventive Action → Trend Verification. Division 23 should not simply produce charts — it should answer whether the AGL system is becoming more reliable, where the weaknesses are, what is recurring, which products are underperforming, and whether corrective action actually worked.
The number of recorded failures during a defined period.
If 18 runway-light failures occur during one month: Monthly failure frequency = 18 failures.
Failure frequency is an absolute count. It should not automatically be interpreted as poor reliability without considering the size of the installed population and operating exposure.
The rate at which failures occur relative to the relevant exposure or population.
Depending on the analysis, exposure may be number of installed units, operating hours, circuit-hours, equipment-hours or maintenance periods.
1,000 fixtures installed, 20 failures occur during a defined period → 20 / 1,000 = 2%. The database should clearly identify the denominator used.
Average operating time between failures for a repairable system/equipment population under the defined calculation method.
A group of CCRs provides 20,000 operating hours and experiences 10 relevant failures → MTBF = 2,000 hours.
The calculation methodology must remain consistent. Do not mix different populations or exposure definitions without clearly identifying them.
Average time required to restore failed equipment to service.
Two systems can have the same failure rate but very different operational reliability if one takes significantly longer to restore.
The period for which an equipment/system function remains unavailable or degraded due to a failure.
The proportion of time equipment is available for its intended function.
or, under corresponding assumptions: Availability ≈ MTBF / (MTBF + MTTR).
Availability should be calculated using a clearly defined measurement period and downtime convention.
A separate, AGL-specific metric: the availability of an AGL series circuit or lighting circuit for its intended operational function.
A single fixture failure and a complete circuit failure have very different operational consequences. Circuit availability therefore provides a more meaningful system-level reliability indicator than fixture-level counts alone.
The proportion of failures involving equipment/components that fail repeatedly after previous repair or replacement.
100 failures recorded, 15 are repeat failures → Repeat-failure rate = 15%.
Define what constitutes a repeat failure — same asset, same component, same failure mode, within a defined period — and keep that definition consistent across the database.
The percentage of maintenance interventions where the failure is successfully corrected during the first intervention without requiring a repeat visit for the same issue.
100 maintenance interventions: 85 resolved first time, 15 required repeat intervention → First-time-fix rate = 85%.
Tracks how the severity distribution changes over time.
| Period | Critical | Major | Moderate | Minor | Degradation |
|---|---|---|---|---|---|
| Q1 | 2 | 8 | 20 | 35 | 12 |
| Q2 | 1 | 6 | 18 | 31 | 15 |
| Q3 | 1 | 4 | 15 | 28 | 22 |
Not simply “are failures increasing?” but also “are the failures becoming more or less serious?”
Track failures month-by-month, from January through December, to identify sudden increases or improvements.
Group data into Q1–Q4. Useful for management review, reliability reporting, supplier review, maintenance planning and project performance review. Quarterly analysis also reduces excessive noise from individual monthly variations.
Compare year over year (e.g. 2026 → 2027 → 2028).
Determine whether overall AGL reliability is improving → stable → deteriorating.
Analyse reliability by equipment category — runway fixtures, taxiway fixtures, CCRs, isolation transformers, PAPI, signs, approach lighting, stop bars, RGLs, ALCMS/control systems.
Compare reliability across airports using failures, failure rate, MTBF, MTTR and availability side by side.
Raw failure counts should not be used alone — a larger airport with thousands of AGL assets will naturally generate more failures than a small airport. Compare normalized metrics where appropriate.
Analyse failures by component — LED module, LED driver, isolation transformer, connector, cable, CCR component, PAPI component, sign component.
This directly supports Division 19 — Product & Component Reliability Failures.
Analyse failure occurrence against environmental conditions — temperature, rainfall, humidity, flooding, sand/dust, salt exposure, lightning, storms, UV exposure.
Heavy rainfall period → increase in moisture ingress, connector failures, transformer failures, base-can water accumulation. This can reveal environmental correlations not visible in normal failure counts.
Compare failures by season or defined operational/environmental period — pre-monsoon, monsoon, post-monsoon, winter, summer. The exact classification can be adapted to the airport location.
Identify whether particular failure modes increase during particular periods.
Analyse failure performance by manufacturer, supplier, product family, model, part number or batch.
Division 19 → Division 23.
Track repeat failures over time.
Q1 → 12 repeat failures · Q2 → 15 · Q3 → 8 · Q4 → 5. A declining trend may indicate successful corrective actions.
Identify physical or functional areas with unusually high failure concentration — a runway section, taxiway, circuit, CCR, base-can area, lighting zone, equipment room, or an entire airport.
If failures repeatedly occur in one runway sector, this chain can reveal an underlying common cause.
Create a ranked list of the most frequently recurring failure modes — e.g. connector high-resistance failure, LED driver failure, moisture ingress, cable insulation deterioration, fixture alignment issue. The actual ranking must come from the database.
Becomes the basis for preventive maintenance priorities, spare planning, product review, engineering improvements and training priorities.
Rank failures based on criticality, safety impact, operational impact, downtime and recurrence — e.g. critical runway lighting failures, PAPI failures, stop-bar failures, multiple circuit failures, major CCR failures, critical control-system failures.
Connects directly to Division 22 — Failure Severity & Criticality.
Create a ranking based on normalized reliability, not simply failure count.
| Rank | Component | Failure Rate | MTBF | Repeat Rate | Criticality |
|---|---|---|---|---|---|
| 1 | Component A | — | — | — | High |
| 2 | Component B | — | — | — | Moderate |
| 3 | Component C | — | — | — | High |
Division 23 is the analytics layer over the entire Failure Database — it does not stand alone. These division-wide references define how the metrics above feed a management reliability dashboard.
Since the source framework specifically targets failure-trend dashboards, Division 23 should ultimately produce a management dashboard built from eight sections:
For senior management, a core KPI dashboard containing the following:
| KPI | Purpose |
|---|---|
| Total Failure Count | Overall failure burden |
| Failure Rate | Normalized reliability |
| MTBF | Reliability |
| MTTR | Maintainability |
| Total Downtime | Operational consequence |
| Equipment Availability | Equipment performance |
| Circuit Availability | AGL system performance |
| Repeat-Failure Rate | Recurring problem indicator |
| First-Time-Fix Rate | Maintenance effectiveness |
| Critical Failure Count | Risk indicator |
| Safety-Critical Failure Count | Safety indicator |
| Top Failure Mode | Priority improvement area |
| Top Unreliable Component | Product priority |
| Top Failure Hotspot | Location priority |
Step 2 — Validate checks for duplicate records, missing dates, missing asset IDs, incorrect classifications, missing downtime and missing failure status.
Step 6 — Trend analyses monthly, quarterly, annual, equipment, component, airport, environment and supplier dimensions.
Step 9 — Verify compares pre-action and post-action reliability.
Data quality is critical for Division 23. Before calculating reliability statistics, ensure the required minimum data is present:
Without these, some metrics can be misleading — 20 failures at Airport A cannot be meaningfully compared with 10 failures at Airport B unless the relative asset population and/or exposure is known.
A good Division 23 analysis should answer five questions:
This prevents management from focusing only on the largest number of failures.
Division 23 is the analytics layer over the entire database:
Division 23 also asks: what keeps failing, where is it happening, which components/products are unreliable, and is reliability improving.
Divisions 01–20 classify individual failures, Division 21 explains why each one happened, and Division 22 rates how serious it was. Division 23 does not add new failure records — it converts the accumulated records into reliability statistics, trends, hotspots and rankings, and feeds a management dashboard that answers whether the AGL system is becoming more reliable, where the weaknesses are, and whether corrective actions actually worked. This makes Division 23 the performance-intelligence layer of the entire AGL Failure Database, and the foundation for predictive reliability, maintenance optimization and risk-based maintenance.
Quick answers about reliability metrics and failure-trend analysis, as covered in Division 23 of the AGL Failure Database.
Division 23 is the reliability statistics and failure-trend intelligence layer. It contains 24 published topics (REL-01 to REL-24) covering failure frequency, failure rate, MTBF, MTTR, downtime, availability, repeat-failure rate, monthly, quarterly and annual trends, hotspot analysis and top unreliable component rankings, and feeds a management-level reliability dashboard.
They fall into four groups: Core Reliability Metrics (10 topics), Monthly, Quarterly & Annual Trend (3), Population & Comparative Analysis (7) and Hotspots & Rankings (4).
No. Divisions 01–20 classify individual failures, Division 21 explains why each happened and Division 22 rates how serious it was. Division 23 converts those accumulated records into reliability statistics, trends, hotspots and rankings.
The core KPI set includes total failure count, failure rate, MTBF, MTTR, total downtime, equipment availability, circuit availability, repeat-failure rate, first-time-fix rate, critical failure count, safety-critical failure count, top failure mode, top unreliable component and top failure hotspot.
The workflow has nine steps: collect data from Divisions 01–20, validate, link RCA (Division 21), apply criticality (Division 22), calculate metrics, trend, identify hotspots, take corrective action and verify improvement. Validation checks for duplicate records, missing dates, missing asset IDs, incorrect classifications, missing downtime and missing failure status.
Data quality is critical. The minimum data includes asset ID, equipment type, failure date, failure mode, operating population, operating period or exposure, repair or return-to-service date, failure severity and root cause where available. Without these, some metrics can be misleading, because 20 failures at one airport cannot be compared with 10 at another unless asset population or exposure is taken into account.
Five: what is failing most often (frequency and rate), what is failing repeatedly (repeat-failure rate and trend), what is failing most seriously (severity and criticality trend), what is taking the longest to restore (MTTR and downtime) and where engineering effort should be focused (hotspot, criticality, recurrence and reliability trend together).
A rising failure trend on one component type usually has an engineering explanation — a batch issue, an environmental factor, an installation practice — real reliability analysis means understanding AGL systems well enough to find it.
Take that knowledge further with Precision AGL Academy.
Learn how to connect:
Develop structured knowledge across AGL fundamentals, components, electrical systems, testing, photometry, troubleshooting, maintenance and commissioning.
Find the failure. Understand the cause. Build the engineering skill.
Please read before relying on any content on this page or on any record downloaded from it.