The bathtub curve — a model of changing failure risk
The bathtub curve describes how failure rate in a population can change with age: initially decreasing, then approximately constant, and later increasing as wear-out becomes more significant. Products need not show identical patterns or clear phase boundaries. Apply the model using evidence for the design and operating environment. NIST: Bathtub curve.
The following phases help organise the assessment; their duration and causes must be established for the population concerned.
Infant mortality
Early failures can arise from manufacturing defects, transport damage or installation errors. Neither a fixed 6–18-month period nor a single dominant cause can be assigned to all street luminaires; investigate the case concerned.
Useful life
A low, roughly constant failure rate due to random events: impact damage, electrical surges, extreme weather, or simply statistical component failures. This is the phase MTBF (Mean Time Between Failures) describes.
Wear-out
Wear-out mechanisms can increase failure risk with age, but the time of an individual failure remains uncertain. Capacitors, seals and LED modules need separate evidence; the curve alone does not identify which component will limit operation.
In practice
- An MTBF calculated under a constant failure-rate assumption is not a prediction of when wear-out starts. Ask which model, conditions and population the figure describes.
- Incoming inspection, correct installation and the manufacturer’s commissioning checks can reduce early-failure risks. Visual checks alone do not re-establish a laboratory IP rating.
- Wear-out risk can be estimated from applicable component models and measured conditions. The uncertainty and variation between units must remain part of maintenance planning.
Root causes — what actually starts the process
Several mechanisms are relevant when assessing a failed luminaire. Their frequency and order depend on design, installation, environment and maintenance; the following list is not a measured ranking of street-light failures.
- Moisture ingress can contribute to failure if water reaches vulnerable conductors or electronics. An IP rating describes the tested configuration and conditions; it does not establish moisture as the cause of a particular failure or guarantee every seal’s future condition.
- Thermal cycling: temperature changes produce differential expansion in joints, circuit boards and housing materials. Fatigue depends on temperature range, number of cycles, geometry and material properties; damage is not inevitable in every installation.
- Corrosion can develop where susceptible metal and exposure conditions permit it. Surface corrosion alone does not prove loss of watertightness. Damage to a sealing surface, joint or housing wall can connect corrosion with moisture ingress; inspect the actual path. TWI: galvanic corrosion.
- Driver component ageing: wet aluminium electrolytic capacitors can lose electrolyte and age faster at higher temperature. Use the selected series’ temperature, ripple-current and life model. A simplified 10 °C rule must stay within its valid range and never justify exceeding a rated temperature. Nichicon, § 2-9.
- LED lumen depreciation is one possible loss of performance; sudden electrical failures can also occur. If maintained lighting no longer meets the project’s applicable EN 13201 criteria, investigate output, optics, dirt, controls and the lighting design even if the luminaires remain lit.
How failure chains develop
A possible failure chain starts with a damaged sealing interface that admits moisture. Condensation can form where a surface is below the dew point. Water and contaminants may then contribute to conductor corrosion or electrical leakage. This is an example to investigate, not a diagnosis: the coolest surface need not be the driver PCB, and a logged “driver failure” does not establish which seal, component or event started the process.
Possible loss of gasket recovery over time; verify material and exposure
Humid outdoor air enters the enclosure through a compromised seal
Dew point reached on the coldest surface in the enclosure
Trace oxidation, electrolytic migration, weakening of solder joints
The three sub-topics that explain failure in depth
The individual failure mechanisms — lumen depreciation, moisture ingress, and driver ageing — each have their own logic and their own timescales. Three articles in this knowledge base go into each in detail.
Service Life · 03
100,000 hours — what it actually means
LM-80 test data, TM-21 projection limits and the documentation needed to assess a long-hour claim.
Read →
Service Life · 06
Moisture, condensation and failure modes
How humid air enters an enclosure, when it condenses, and what that means for the electronics inside.
Read →
Service Life · 08
How driver capacitors affect service life
The Arrhenius principle, electrolytic capacitor ageing, and why driver temperature management matters more than headline MTBF.
Read →
Procurement implications
For procurement, request evidence for LED lumen maintenance, the selected driver’s life at stated temperatures, seal materials and the complete construction. Distinguish a constant-rate MTBF calculation from wear-out estimates, and identify the test object, conditions and failure criterion behind every hour figure.
Combine this evidence with the specified IP and IK ratings and an assessment of the site’s corrosivity. An environmental category is not itself a luminaire certification. The combined evidence is more informative than purchase price and luminous efficacy alone.
Summary
Failure risk can be assessed with component evidence and operating conditions, though the time of an individual failure remains uncertain. Early failures, approximately constant-rate failures and wear-out are useful population models. Request evidence for drivers, seals, LED modules and the complete offered configuration.
Sources and further reading
- NIST Engineering Statistics Handbook — Bathtub Curve
- Nichicon — Application Guidelines for Aluminum Electrolytic Capacitors, sections 2-1 and 2-9
- IEC 60529:1989+AMD1:1999+AMD2:2013 CSV — Degrees of protection provided by enclosures (IP Code)
- TWI — What is Galvanic Corrosion and How Can it be Prevented?
- Parker — O-Ring Handbook