In this short article, we want to highlight the underlying mechanisms which lead to persistent imbalances in large BESS systems.
Key Takeaways
The BMS relies on specific conditions (like high SoC) to perform balancing—these are not always met.
Large SoC differences between cells can block balancing entirely.
Balancing is a slow process, making it ineffective for large or recurring imbalances.
Self-discharging cells can cancel out balancing effects.
Persistent imbalances may require hardware replacement.
Recurring patterns in the data suggest deeper degradation issues.
1. Balancing Basics: How It’s Supposed to Work
The BMS performs cell balancing by bringing all cells to a minimum SoC—typically above 80%—and then gradually discharging the cells with the highest voltages to match the rest.
This method works best when all cells are close at a similar state of charge level. Some systems can also trigger balancing at very low SoC levels, such as below 20%, though this is less common.
Passive balancing techniques like these are intended to smooth out small differences. However, they struggle when large imbalances are present.
2. Large Imbalances Prevent Balancing
Balancing cannot begin unless every cell in a string is within the allowed voltage range and above the balancing SoC or voltage threshold. For example, if one cell is at 70% SoC and another is already at 100% (representing an SoC imbalance of 30%), the system cannot charge the entire string further without overcharging the full cell.
This leaves the low cell permanently below the threshold, and balancing will never start. LFP cells are especially prone to this issue because their flat voltage profile makes SoC estimation more difficult. In such cases, standard balancing logic is ineffective.
3. Self-Discharge Cancels Out Balancing
Even if the BMS successfully initiates balancing, some cells may exhibit higher self-discharge rates than others. This internal leakage gradually reduces their voltage, pulling them back below the required threshold.
As a result, the BMS halts the balancing process, unable to complete it. The problem then repeats, creating a loop of imbalance and interruption. Over time, this leads to chronic performance degradation.
4. Balancing is Extremely Slow
Balancing proceeds at a very gradual pace—typically around 7 to 10 millivolts per day for NMC systems in voltage spread reduction. That translates to only about 1 to 2% of SoC per day. If a string has a 30% SoC imbalance, it could take several weeks of uninterrupted balancing to fully correct it.
In real-world operation, interruptions are common, making full balancing virtually impossible. This is why large imbalances often persist over time.
5. Balancing Interrupts Revenue Operations
To trigger balancing, a battery system must be brought to a high or low SoC and held there for extended periods. During these times, the system cannot participate in regular energy trading or grid services.
This idle period results in direct financial losses for asset owners and operators. Because of this, balancing cycles are typically scheduled occasionally or avoided entirely. The result is a trade-off between long-term performance and immediate revenue generation.
6. Hardware Replacement May Be More Effective
When imbalances become persistent and resist correction, physical maintenance may become the only viable solution. Replacing the affected modules or strings can be quicker and more cost-effective than attempting extended balancing cycles.
This approach is especially common when specific components show signs of degradation or recurring faults. It also helps restore system availability faster than balancing alone. For operators, it’s a matter of cost-benefit analysis.
7. Recurring Patterns Signal Deeper Issues
Operational data often shows imbalances that appear to resolve temporarily, only to re-emerge days or weeks later. This indicates that while the BMS may temporarily balance the cells, underlying degradation is causing them to drift again. In these cases, the root cause lies deeper than the balancing logic—it may be different degradation or performance state or manufacturing variance.
Persistent imbalance patterns should therefore be treated as diagnostic signals, not just operational nuisances. Identifying these early can prevent long-term capacity loss or safety issues.