An accelerated aging trial can report acceptable retained breaking strength while the joint has already moved enough to disturb its function. It can also produce a failure mechanism that would not occur under the intended service conditions. Define both the endpoint and the mechanism before interpreting test duration.
For a bonded transformer packet, useful endpoints may include relative slip, opening, loss of bond continuity, damage to the original coating, or a change in liquid compatibility. Their relevance depends on the construction and demand. A single strength percentage is not a complete description of aging.
Translate function into an observable endpoint
Choose observations that the test can measure reproducibly. If geometry must remain stable, record movement under the selected sustained load and after unloading. If insulation is at risk, include the appropriate electrical response. If exposure includes insulating liquid, examine material and liquid changes with controlled blanks.
| Functional endpoint | Observation | Possible failure missed by strength alone |
|---|---|---|
| Positional stability | Slip and residual movement | Time-dependent distortion |
| Bond continuity | Mapped disbond or crack growth | Local failure diluted in a global test |
| Coating preservation | Failure plane and electrical comparison | Original insulation damaged beneath adhesive |
| Liquid compatibility | Selected chemical and dielectric changes | Extractables without visible joint failure |
Acceptance criteria should come from the intended function and qualified measurement, rather than from the easiest quantity to test.
Check whether acceleration changes the mechanism
Compare fracture location, material condition and deformation pattern across the tested exposures. A higher-temperature condition may alter the adhesive state or introduce reactions absent from lower-temperature conditions. If the mechanism changes, fitting all results to one lifetime relationship can be misleading.
Keep load, atmosphere, liquid identity, specimen geometry and cure history documented. Temperature alone cannot represent every relevant exposure. A dry, unloaded coupon may reveal one aging process while omitting joint creep or liquid-mediated changes.
Preserve time-resolved evidence
Use retained specimens or nondestructive observations at justified intervals. An endpoint test shows what remains at one time; a sequence helps distinguish gradual deterioration from an abrupt change. Where destructive tests are used, specimen-to-specimen variation needs to be considered in interpreting the trend.
Record censored outcomes honestly: a specimen that has not failed when the trial ends has survived that trial, not an unlimited service period. A specimen removed for another reason has a different information status. These distinctions matter if the results will support a later statistical model.
Separate adhesive-system life from whole-core life. The core includes other interfaces and functions, and service conditions may vary through time. Even a defensible material acceleration relationship does not automatically predict the operating life of the assembled transformer.
The trial should produce an endpoint map and an assessment of mechanism consistency. Only after those are established is it reasonable to examine whether a supported acceleration model applies. High-temperature test hours, by themselves, are a process record rather than a conversion table to service years.
Technical references
- voestalpine. Backlack: Lamination bonding.
- voestalpine. Backlack: Testing methods and production of test specimens.