Battery Cell Balancing via Model-Based State of Charge Estimation

The patent replaces direct voltage measurement-based SOC estimation with a model-based estimation system using equivalent circuit models and Kalman filtering. This substitutes the direct measurement approach (analogous to mechanical measurement) with a computational model that processes multiple signals (voltage, current, temperature) to estimate SOC, thereby improving accuracy while maintaining system reliability through model-based prediction.

The patent introduces model parameter estimates as an intermediary between raw voltage measurements and final SOC calculation. The Kalman filter acts as a mediator that processes noisy voltage and current measurements through the equivalent circuit model to produce accurate SOC estimates, resolving the contradiction between measurement precision and reliability.

The patent implements dynamic excitation strategies that adaptively adjust the excitation signal based on current operating conditions. The system dynamically determines when and how to apply excitation to satisfy persistent excitation conditions for accurate parameter estimation, while coordinating with the vehicle control system to maintain required acceleration performance, thus resolving the contradiction between estimation accuracy and vehicle speed requirements.

The patent employs periodic excitation signals applied at specific intervals when vehicle acceleration requirements are satisfied. The excitation is applied periodically to gather data for parameter estimation, but only during periods when it does not conflict with vehicle acceleration needs, balancing measurement precision with speed requirements.

The patent implements a feedback-based cell balancing control system that continuously monitors SOC estimates and cell voltage differences. The balancing operation is activated only when the estimated SOC difference exceeds a threshold, and deactivated when balance is achieved. This feedback mechanism maintains cell balance stability while avoiding unnecessary balancing operations that would waste energy.

The system uses model-based SOC estimation to self-determine when cell balancing is needed, eliminating the need for continuous monitoring and manual intervention. The controller automatically adjusts balancing operations based on real-time estimates, maintaining cell balance stability while minimizing energy consumption by performing balancing only when necessary.