Accurate state-of-charge estimation is essential for battery management systems in grid-scale energy storage and electric vehicles, yet lithium iron phosphate batteries present a unique technical challenge. Unlike other lithium chemistries, LFP cells exhibit a characteristically wide plateau in their open-circuit voltage curve—a region where voltage changes minimally despite significant variations in charge state. This flat response makes traditional voltage-based estimation methods unreliable.
Researchers have proposed a region-dependent estimation framework that adapts its approach based on the battery's operating zone. In the voltage-insensitive plateau region, the method relies on ampere-hour integration, which accumulates charge flow directly without depending on voltage signals. Once the battery enters the voltage-sensitive steep region, an adaptive extended Kalman filter activates to perform closed-loop correction using voltage measurements as feedback.
The strategy incorporates online parameter identification through an adaptive forgetting-factor recursive least-squares algorithm, enabling the system to track dynamic changes in battery impedance as operating conditions evolve. Under high-excitation scenarios with sufficient current variation, parameters update continuously. During constant-current charging or discharging phases with limited excitation, the system reverts to offline-identified parameters to avoid unreliable updates.
Validation testing on 280-ampere-hour LFP cells demonstrated maximum absolute state-of-charge errors below 3% across multiple test profiles, including constant-current, dynamic stress, and hybrid pulse power characterization tests. The method showed improved robustness to initial SOC uncertainty, sensor noise, and parameter variations—practical concerns in deployed systems.
For energy storage operators and battery management developers, this approach offers a path toward more reliable monitoring of large-format LFP packs used in stationary grid storage and heavy-duty electric vehicles, where estimation accuracy directly impacts system efficiency, lifespan prediction, and operational safety.
