Battery Ageing, Swelling Evolution & Preload Retention

Ageing Changes the Mechanical State

Battery mechanical state evolves even when the external load environment is unchanged. Cells can exhibit irreversible growth, pads can relax or take compression set, adhesives can creep and polymer frames can change stiffness. The module force that was correct at build may therefore move significantly over life. Structural–thermal analysis should treat beginning-of-life, intermediate and end-of-life states as distinct configurations rather than assume one permanent preload.

Separate Reversible and Irreversible Contributions

Cell thickness change can contain reversible state-of-charge breathing, thermal expansion and irreversible ageing growth. These contributions occur on different timescales and should be separated where data permit. A model that treats the total observed change as instantaneous swelling can overpredict dynamic force, while a model that ignores long-term growth can underpredict end-of-life enclosure load. The history used for preload assessment should reflect the actual sequence of assembly, cycling, storage and operation.

Closed-Loop Stiffness Controls Force Evolution

The change in stack force is governed by the stiffness of the cells, pads, end plates, ties and surrounding structure. A compliant retention feature can accommodate growth with limited force rise; a stiff frame converts small growth into large load. A reduced spring model remains a powerful independent check on detailed contact FEA.

Approximate force change:  ΔF ≈ k_system Δ_growth

where k_system is the effective stiffness of the complete closed compression loop, not merely the cell stiffness.

Relaxation Competes With Growth

Pads, elastomers, polymers and adhesives can lose force with time and temperature while cells grow. The net preload is the result of both processes. Analyses should avoid simply adding a growth force to the original assembly force if the compliant elements have relaxed. Where detailed viscoelastic models are unjustified, bounded retained-stiffness or retained-force cases can provide a transparent engineering envelope.

Thermal Contact Can Move With Preload

As module compression changes, thermal interface contact pressure and thickness can change. That can alter conductance and therefore cell temperature, which may feed back into ageing and swelling. A fully coupled life model is rarely practical, but targeted sensitivity studies can determine whether this feedback is strong enough to matter. If thermal performance is insensitive across the plausible pressure range, a simpler one-way life assessment is defensible.

End Plates, Ties and Enclosure Features See Life-Cycle Load

Ageing-induced force can increase end-plate bending, tie-rod load, weld force and enclosure reaction. The most highly stressed condition may therefore occur late in life rather than during initial assembly. Conversely, relaxation can reduce restraint and permit larger cell motion under vibration or crash. Both high-force and low-force life states should be considered where they govern different failure modes.

Data Pedigree Is Often the Main Uncertainty

Growth and stiffness data vary with chemistry, format, temperature, state of charge, cycling history and manufacturing population. Supplier curves or limited coupon/cell tests should be tied to the assessed cell revision and duty. Extrapolation beyond the tested life should be visible. It is usually better to present bounded life-state scenarios than a falsely precise single ageing trajectory.

Inspection and Service Evidence Can Update the Model

Where accessible, module thickness, pack flatness, tie load, cell pressure indicators or teardown measurements can provide evidence of actual life-state progression. The model can then be updated using measured geometry or retained preload. This creates a stronger basis for continued-service decisions than relying indefinitely on the original design assumption.

Verification

Check spring-network predictions against FEA force change, confirm that total reaction equals stack force, inspect contact status through the full life sequence and verify that material curves use the correct time/temperature state. Sensitivity should include growth magnitude, restraint stiffness and relaxation because those three variables often dominate end-of-life force.

Population Scatter and Cell Replacement

Ageing growth is not only a mean-value problem. Cell-to-cell variation can create local force concentrations within a nominally uniform module, especially when replacement cells or mixed production lots are present. A practical structural assessment should consider the spatial pattern of plausible growth, not just uniform expansion of every cell. Bounding cases such as one locally overgrown cell, an end-cell growth bias or a module-wide upper-percentile state can reveal whether end plates and compliant layers redistribute the variation safely.

Life-State Selection Should Follow Failure Mode

Different failure modes can govern at different points in life. Maximum end-plate stress may occur at high accumulated growth; vibration slip may be worst after preload relaxation; seal or thermal contact can degrade at an intermediate state. The qualification basis should therefore identify the life state associated with each critical requirement rather than declaring one universal end-of-life model. This also clarifies which aged tests or measurements are needed to validate each prediction.

Engineering Outcome

A defensible ageing assessment describes how the battery’s mechanical state evolves through life and shows that high-preload and low-preload consequences are both controlled. It connects cell evidence, structural restraint and interface behaviour to end-of-life margins rather than treating ageing as a generic knock-down factor.

End of life is not automatically the maximum-force state: cell growth, pad relaxation and structural creep compete. Model the closed mechanical loop.

Key takeaways

• Represent ageing as an evolving assembly state, not a single property reduction.
• Separate reversible breathing from irreversible growth where evidence permits.
• Bound the competing effects of growth, relaxation and restraint stiffness.