State of Health: Measuring BESS Degradation Before It's Too Late
Every battery storage system is degrading. The real question is whether you are measuring it or finding out too late. State of Health determines deliverable capacity, which modules need replacing, and what the asset is worth to lenders and off-takers.
Every battery storage system is degrading. The real question: are you measuring it — or finding out too late?
What State of Health actually tells you
State of Health (SOH) tells you how much capacity your BESS can actually deliver, which modules need replacing, and what your system is worth to lenders and off-takers at any point in its life.
That last point is the one most often overlooked. SOH is not just an engineering metric — it is the input to warranty claims, augmentation planning, refinancing, and the residual value line in any transaction involving the asset.
Why most systems cannot measure it properly
Most operational BESS systems lack a reliable way to track SOH without interrupting operations. The textbook method is a full controlled charge–discharge cycle against a known reference — which means taking a revenue-generating asset offline, on a schedule, for the privilege of learning what it can still do.
In practice that test rarely happens. Operators fall back on the BMS's own estimate, which is typically a coarse model calibrated at the factory and increasingly wrong as the pack diverges from its assumptions.
Measuring without stopping the asset
The alternative is to infer SOH from operational data the system already produces — voltage and current behaviour across normal cycles, module-level divergence, thermal response, and how the pack behaves at the edges of its operating window.
Done continuously, that yields something a periodic capacity test never can: a trend. Knowing a pack is at 91% matters far less than knowing it lost four points in eight months while its siblings lost one.
Why it compounds
Degradation is not uniform. A handful of weak modules set the usable capacity of the whole string, and they get worse faster under the same duty. Catching divergence early is the difference between rebalancing at the control layer and replacing hardware.
Opens the original post on LinkedIn — like, comment, or share it there.