During operation, lithium-ion batteries experience a mechanical deformation due to the strain induced by the electrochemical processes, involving the interaction of lithium ions with the crystal structure of the active material of the electrodes. Such deformation originating at the atomic scale has an impact on the macroscopic structural deformation of the battery. Measurements of this macroscopic deformation can be used to diagnose battery states, such as degradation mechanisms and state of health, and to improve state of charge estimation. This approach is enabled by the strong correlation between the lithiation of the electrodes and the macroscopic deformation of the battery, which is significantly stronger than the correlation with voltage, the traditional parameter used in battery control and diagnostic algorithms.
In terms of battery diagnostics, this work identifies the critical features of differential mechanical measurements—specifically the second derivative of expansion and incremental expansion—and explains, for the first time, the correlation between mechanical and voltage responses for common active materials (NMC, LFP, and LCO). Furthermore, it demonstrates that differential mechanical curves can be used similarly to differential voltage and incremental capacity curves, with the crucial advantage that the key features of the mechanical curves remain detectable even at high current rates, unlike voltage-based methods which typically require very low currents (C/20-C/10). This makes mechanical measurements a promising tool for battery diagnostics in real-world applications outside the laboratory. Accordingly, the authors propose a procedure to estimate stoichiometric limits, electrode capacities, and battery health using differential mechanical measurements. Results show that the estimation made at low current (C/20) using voltage-based methods align closely with those obtained at high current (1C) using the deformation-based methodology developed in this work. Additionally, this methodology enables battery health estimation from partial recharges, making it highly suitable for real-world automotive applications.
Concerning the state of charge estimation, the macroscopic battery deformation gives a direct information of the electrode lithiation independently on the current applied, as demonstrated mathematically and observed experimentally in previous works. These features makes deformation-based methods more reliable, precise and easier to implement compared to voltage based methods. An algorithm based on the Extended Kalman filter principle is set up to get the state of charge estimation. The algorithm is flexible and can be run with voltage or deformation measurements, independently or in combination. The results indicate that including deformation measurements can halve the error in estimating the charge state, with the error difference between voltage- and deformation-based algorithms growing as the battery ages.