Cell Divergence in Grid-Scale BESS: How Ingro EMS Recovered 10.7% Discharge

The degradation that eventually breaks a grid-scale BESS doesn't start with ageing. It starts at the factory.

No two cells leave the production line identical. Even from the same batch, the same day and the same manufacturer, every cell has slightly different internal resistance and capacity. The differences are small and invisible at first. But in a series-connected string, they compound with every charge-discharge cycle — silently and continuously.

The divergence cliff

In the early stages the system still looks healthy. State of health reads fine and operators see nothing alarming — until the pack crosses what the industry calls the divergence cliff. The curve that looked linear suddenly breaks. Not because something new went wrong, but because something that was always there finally became unmanageable.

Ingro EMS is built to catch this pattern before the threshold is crossed, while there is still room to act. It monitors every module's maximum and minimum cell voltage continuously, flags inter-cell divergence as it builds, and intervenes at the control layer before the OEM BMS is forced into a hard cutoff. The result is not just fault prevention, but recovery of capacity the system was already losing.

Results at one of India's earliest grid-scale BESS deployments

Before intervention, Ingro EMS logged:

• Inter-cell voltage spread exceeding 7x the healthy threshold
• An average of 3.5 voltage faults per week — over-voltage, under-voltage, cell divergence, and inter- and intra-module divergence
• DC bus voltage spikes beyond safe limits, caused by abrupt BMS cutoffs mid-operation

After deployment:

• Nearly zero faults over 4 consecutive weeks
• Discharge energy per cycle up 10.7% (63.19 kWh under hard step and cutoff vs 69.99 kWh with Ingro EMS linear derating)
• 9.2% recovery in usable SoC range
• Zero hardware replaced

A software intervention at the EMS layer controls the problem; it does not erase it. The divergence is managed, not cured. What changes is that it is now visible, bounded and no longer compounding unchecked.

The real cost of module replacement

The conventional response to cell divergence is module replacement. Beyond the rupees, it costs:

• Supply chain wait: lead times in India are months, not days
• Recommissioning downtime: every rack offline is lost revenue
• No guarantee: new cells enter the same unmanaged conditions, and the divergence starts again
• No learning: hardware replacement generates zero operational intelligence
• Wasted assets: recoverable cells get discarded

Cell divergence is not a hardware problem waiting for a hardware solution. It is a visibility and control problem — and that is precisely what Ingro EMS is built to solve.