The core KPIs
| KPI | What it tells you | How it's usually calculated |
|---|---|---|
| State of charge (SoC) | How full the battery is right now | Stored energy ÷ usable capacity, reported by the BMS |
| State of health (SoH) | How much capacity is left compared with new | Measured usable capacity ÷ rated capacity at beginning of life |
| Round-trip efficiency (RTE) | How much charged energy comes back out | Energy discharged ÷ energy charged, same boundary and period |
| Availability | Whether the system could perform when needed | Available time (or capacity) ÷ total time (or capacity) |
| Equivalent full cycles (EFC) | How hard the battery has been used | Total energy discharged ÷ usable capacity |
| Dispatch adherence | Whether it delivered what was scheduled | Delivered energy or power ÷ scheduled, per window |
| Auxiliary consumption | Energy used by HVAC, BMS and controls | Metered auxiliary load over the period |
Why measurement boundaries matter
The same battery can report very different numbers depending on where you measure:
- DC side (at the battery): excludes inverter (PCS) and transformer losses. RTE looks high.
- AC side (at the PCS output): includes conversion losses.
- Point of connection (site meter): includes transformers and often auxiliary load. This is what commercial contracts usually use, and it is the lowest RTE.
Always state the boundary next to the number, and don't compare an AC-side RTE from one site with a DC-side figure from another.
Worked example
Worked example · illustrative numbersRTE and cycles for a 40 MW / 80 MWh system
Over a month at the AC meter, the system charged 2,300 MWh and discharged 2,010 MWh.
RTE = 2,010 ÷ 2,300 = 87.4% (AC-side).
Equivalent full cycles = 2,010 ÷ 80 = 25.1 cycles in the month, a little under one a day.
If one rack is derated by the BMS for thermal reasons, the evening discharge falls short of schedule and the dispatch adherence KPI shows it before the monthly RTE does.
Understanding state of health
SoH is the KPI most often misread. Two things to keep in mind:
- BMS-reported SoH is an estimate. It comes from the manufacturer's model. A periodic capacity test, a full charge and discharge at a defined rate, gives the measured figure.
- Degradation is not linear. It depends on temperature, depth of discharge, C-rate and time. A container running hotter than its neighbours will usually fade faster.
- Use it for planning. Comparing SoH across containers and over time is what drives augmentation decisions: when and how much capacity to add to keep meeting contracted output.
What to watch day to day
- Cell temperature spread within a rack or container. A growing spread is an early warning of cooling or cell problems.
- Cell voltage imbalance, and how often balancing runs.
- Derating events: when and why the BMS limited power.
- Scheduled versus delivered energy in each dispatch window.
- PCS and communication alarms that make the system unavailable.
How nxtGrid helps
How nxtGrid helps
- Track SoC, SoH and availability at the levels your BMS and meters provide
- Exception views for temperature spread, imbalance, derating and alarms
- Planned versus measured dispatch with clearly stated metering points
- Route storage issues into work orders like any other asset