Take a paperclip and bend it a few degrees back and forth. It can survive a surprisingly long time. Now bend it far, and it snaps after a handful of tries. Engineers know this as fatigue: how much you bend each time often matters more than how fast you do it. A recent multiphysics study on vibrating lithium-ion cells reaches a similar verdict, and the authors call it amplitude dominance.
Two knobs, two different roles
Vibration has two main settings. Frequency says how many times per second the cell is shaken. Amplitude says how far it moves each time. In the model, both feed the same base excitation, y(t) = A sin(2πft), and the acceleration the cell feels is proportional to A·f². So each knob raises the load, but through different routes: frequency also adds more cycles per unit of time, while amplitude raises the size of each cycle.
What the study reports
The paper’s conclusions can be summarised in a few points:
- Stress rises with both. Peak von Mises stress climbs with amplitude (reaching roughly 32 MPa across the 0.2 to 0.6 range) and with frequency (approaching about 24 MPa across 50 to 200 Hz). With no vibration, stress is negligible.
- Capacity fade and resistive heating respond more strongly to amplitude. The authors say larger displacement drives nonlinear stress accumulation.
- Frequency mainly changes the rate. Higher frequencies speed up early-stage resistance growth, but the overall effect on resistance is smaller than that of amplitude.
- SEI growth rises with both. Thicker films appear at higher amplitude and at higher frequency.
- Heating follows the same pattern. Higher amplitudes give a larger temperature rise, and the contour maps over 2,000 cycles show warmer regions as amplitude and cycle count increase.
Why the fatigue picture fits
The authors tie this to high-cycle fatigue. Crack growth in the model follows a power law of stress, so a small increase in stress per cycle produces a much larger increase in damage rate. Frequency, by comparison, mostly changes how quickly cycles accumulate. Large stress swings open cracks, break binder contacts and delaminate coatings, which in turn expose fresh surface for SEI formation, raise resistance and add Joule heat.
Larger displacement also increases micro-slip and friction inside the electrode, so more vibration energy ends up as heat and interface damage.
A quick illustration
Using my own numbers rather than the paper’s: suppose damage per cycle scales as stress cubed. If amplitude doubles, stress per cycle doubles, and damage per cycle rises eightfold. If frequency doubles instead, the cell receives twice as many cycles, and, ignoring the extra acceleration for the moment, damage rises about twofold over the same time. This is the intuition behind “how far matters more than how often”, although in the paper’s own equations frequency also raises acceleration through the f² term, so it is not a perfectly clean split.
The feedback loop makes it worse
Higher amplitude does not only damage the structure. Rising resistance generates more heat, and warmer cells speed up SEI growth and electrolyte decomposition. That in turn raises resistance again. The authors describe this as a positive feedback loop, and they argue it makes amplitude’s influence grow with cycle number.
Reading it with care
- Ranges differ. Amplitude was varied over about 0.2 to 0.6, while frequency spanned 20 to 200 Hz (or 50 to 200 Hz in some figures). A fourfold or tenfold change in one is not the same size of change as a threefold change in the other, so the ranking depends on the ranges tested.
- Units are unclear. The paper gives amplitude values without making the unit obvious, which makes it hard to reproduce the comparison exactly.
- It is a simulation. Several damage parameters are phenomenological, so the ranking is a modelled trend, not an experimental measurement.
- Real roads are not sine waves. Random vibration, shocks and resonance can behave differently, and the authors note that a controlled single-tone study is a foundation for those more complex cases.
What it means for design and testing
If amplitude is the main lever, the priority is limiting how far the cell moves. Stiffer mounts, damping layers, isolation between pack and chassis, and avoiding resonance bands that amplify motion all reduce displacement at the cell. For test planning, it also suggests that qualification programmes should not only vary frequency sweeps but also deliberately include the highest realistic amplitudes. A cell that survives many gentle shakes may still fail under a few large ones.
