A battery energy storage system (BESS) rarely fails dramatically. Far more often it underperforms quietly: a little less energy back per cycle, a converter that trips now and then, cooling that eats more power in May than anyone modelled. The dashboard still shows green.
The way to catch that is a short list of key performance indicators (KPIs), measured consistently and compared against what the contract or the business case assumed.
The five numbers that matter
1. Availability
Availability asks whether the system was ready to operate when it was supposed to be. On a capacity contract it is the product being sold, which is why we gave it a full article on availability guarantees and spare parts.
SECI’s standalone storage tenders — for example its 500 MW / 1,000 MWh and 50 MW / 100 MWh tenders — have asked for a minimum of 95% availability on an annual basis. Shortfalls have been priced steeply: liquidated damages at twice the monthly capacity charge for the capacity that was unavailable.
One analysis of an early SECI tender pointed out that availability was measured only during scheduled time blocks. That detail matters: a system that is down overnight when nobody schedules it may still score well.
2. Round-trip efficiency (RTE)
RTE is the energy you get back out divided by the energy you put in, expressed as a percentage. Every cycle loses some energy as heat in the cells, the power conversion system and the transformer.
The common tender floor is 85% AC-to-AC, measured monthly. Penalties vary by tender. SECI’s 50 MW / 100 MWh tender set liquidated damages of ₹3.50 per kWh of excess conversion losses below 85%, and withholding of the whole monthly payment if RTE drops below 70%. Another SECI clarification used ₹2 per unit for RTE between 70% and 85%. Some tenders also reward performance above 85%.
The measurement boundary is the thing to check. One tender stated its RTE figure excluded auxiliary power. Two systems both claiming 85% can be very different if one counts cooling and the other does not. Our explainer on round-trip efficiency and degradation covers where the losses come from.
3. Usable capacity (state of health)
Lithium iron phosphate cells lose capacity slowly with cycling and calendar age. The battery management system (BMS) reports a state-of-health estimate, but an estimate is not a measurement.
The reliable approach is a periodic capacity test: a full, controlled discharge at rated power, compared against the baseline energy recorded at site acceptance testing and against the curve in the warranty and degradation guarantee. Between tests, trend the energy discharged per full cycle from your monitoring data. A steady drift faster than the warranty curve is an early warning that augmentation or a warranty claim may be coming.
4. Auxiliary consumption
Cooling, fire detection, controls and lighting draw power whether the battery is cycling or not. In Indian summers, heating, ventilation and air conditioning (HVAC) can be a meaningful share of losses. Track auxiliary energy separately, by month, so a cooling fault or a mis-set temperature setpoint shows up as a number rather than a mystery. See auxiliary consumption and losses for typical sources.
5. Dispatch response
Did the system deliver the power and energy it was scheduled to, when it was scheduled? Track scheduled versus actual megawatt-hours per time block, and count trips, faults and partial deratings. For a grid asset this links directly to deviation charges; for a factory it decides whether a demand-charge peak was actually shaved.
Where the data comes from
None of this works without good data. The energy management system (EMS), BMS and supervisory control and data acquisition (SCADA) layer must log energy in and out at the AC meter, auxiliary energy on a separate meter, state of charge, temperatures and fault events — time-stamped and kept. Our guide to SCADA and remote monitoring explains how that stack fits together.
A simple monthly KPI sheet with these five lines, each against its target, is worth more than a dozen live dashboards nobody reviews.
Common mistakes
- Comparing RTE figures with different boundaries. DC-to-DC cell efficiency is always higher than AC-to-AC system efficiency, which is higher than AC-to-AC including auxiliaries.
- Trusting the state-of-health display without testing. The BMS estimate can drift; a controlled discharge test is the ground truth.
- Averaging away problems. An annual availability figure can hide a bad month. Look at monthly numbers even if the contract is annual.
- Ignoring partial deratings. A system running at 70% of rated power is not “down”, but it is not delivering what you paid for either.
What this means for you
- If you own or are developing a grid-scale project: build the KPI definitions — especially the RTE boundary and the availability measurement window — into your supply and O&M contracts so they mirror your offtake contract. Penalties in SECI standalone storage tenders flow straight to your margin.
- If you are a commercial or industrial buyer: ask your supplier which of these five KPIs they will report monthly, and how. A C&I cabinet system used for peak shaving earns only if it is available and efficient at the peaks.
- If you are comparing quotes: ask each bidder for guaranteed RTE with a stated boundary, guaranteed availability with a stated window, and the capacity test procedure.
Tender performance requirements and penalty formulas change from one tender to the next and by notification; verify the current terms in the specific request for selection and its clarifications. To see how these numbers translate into savings for your site, try our BESS savings calculator or talk to our team.