A packet with a local disbond may have almost the same first resonant frequency as an intact packet. That does not make the bond sound. The selected mode may transfer little strain through the damaged interface, or the local stiffness loss may be too small relative to the packet’s overall response.
Resonance screening answers a structural question about the modes actually excited and observed. To use it as a bonding check, determine how those modes respond to defects at different positions. A single frequency comparison lacks that spatial information.
Find where the mode stores strain energy
In a simplified linear model, a mode depends on the distribution of stiffness and mass. Reducing stiffness where that mode develops substantial relative deformation is more likely to change its response than reducing stiffness in a nearly rigidly moving region. This is a sensitivity argument, not a predicted percentage shift.
Map relative motion or calculate a justified modal strain-energy distribution for the intact specimen. Pay particular attention to interfaces where the bond transfers shear or opening. Large external displacement at a location does not necessarily mean high strain there; an entire region can move substantially as a nearly rigid body.
| Defect location in relation to a selected mode | Likely screening difficulty | Additional comparison |
|---|---|---|
| Strong interlayer deformation | Potentially greater modal sensitivity | Known local disbond at that region |
| Nearly rigid motion | Frequency may change little | Another mode with different strain distribution |
| Near support or fixture contact | Fixture changes may imitate damage | Repeated mounting on intact references |
| Small isolated interface region | Global response may dilute the effect | Local bond inspection |
Confirm these hypotheses with representative defects. The prediction should guide the comparison rather than serve as proof of detection.
Control the boundary that defines the measurement
Support positions, contact stiffness and preload can shift resonance or change damping. Reproduce the intended screening fixture and measure intact specimens after repeated remounting. Use that distribution to understand how large a response difference can arise without a bond defect.
Record excitation position and response-sensor position. A mode may be absent from the measured spectrum because the actuator or sensor lies near a node. A missing peak is not itself evidence of bond failure. Compare mode shapes where feasible, instead of matching peaks by frequency order alone.
Keep frequency, damping and amplitude separate
A disbond can affect local friction and energy dissipation without causing an easily measured frequency shift. Peak amplitude also depends on excitation force, sensor coupling and damping. Normalize or otherwise account for these measurement conditions before interpreting an amplitude change as stiffness loss.
A practical screening report lists the modes used, their identified shapes, repeatability and demonstrated defect sensitivity. Mark the regions that none of the selected modes interrogates effectively. Those regions need another method if their bond integrity matters to the design.
The resulting decision is specific: resonance can screen a validated defect class in defined locations and mounting conditions. A packet should not be accepted solely because one peak remains near its nominal position. The sensitivity map explains exactly which local damage that observation may fail to reveal.
Technical references
- voestalpine. Backlack: Lamination bonding.
- voestalpine. Backlack: Testing methods and production of test specimens.