1,000 kWh BESS Fails Full-Power Export and Ride-Through Tests
A 500 kW / 1,000 kWh BESS is commissioned behind a 750 kVA, 12.47 kV/480 V transformer with 5.75% impedance on a utility feeder regulated to 1.03 pu at the primary. On the first full-power export test at unity power factor the inverters trip on overvoltage after 40 s at roughly 94% output, and the point-of-common-coupling log shows 1.09 pu. A separate run confirms correct anti-islanding operation. During a subsequent 0.5 s voltage sag to 0.70 pu the system trips offline, which the interconnection agreement prohibits; the agreement requires ride-through of 0.5 s sags to 0.70 pu and volt-VAR regulation within ±0.02 pu of setpoint. The capacity test then delivers 926 kWh at the DC terminals over 2 h 4 min, and afterward the site's SoC display reads 4% lower than the measured value. Nameplate is 1,000 kWh DC, and the vendor's beginning-of-life tolerance is minus 5%.
1. Estimate the voltage rise causing the overvoltage trip and recommend corrective actions in priority order.
Read the worked answer
Estimate. For a transformer at unity power factor, the approximate rise from exporting S through a transformer of rating S_t and impedance Z% is ΔV ≈ Z% × (S/S_t) = 5.75% × (500/750) = 3.83%. Starting from 1.03 pu at the primary gives about 1.068 pu, and with feeder drop, tap error, and instrument accuracy that matches the logged 1.09 pu. The inverters trip because the overvoltage element (commonly around 1.10 pu with a defined delay) is reached before full output.
Actions, in cost order:
- Verify the transformer tap. If it is set to boost the secondary, dropping one tap (about 2.5%) buys roughly 0.025 pu immediately.
- Enable volt-VAR. Absorbing reactive power lowers the PCC voltage; set the deadband and droop so regulation stays within the utility's ±0.02 pu band.
- Cap export at 470–480 kW or curtail on a voltage feedback signal. It costs revenue but works immediately.
- Re-verify the utility primary voltage and the transformer nameplate impedance/X-R; if the primary really sits at 1.03 pu, coordinate settings with the utility.
Do not simply raise the overvoltage trip point - that breaks the interconnection agreement and the ride-through/overvoltage requirements.
2. The system trips during a 0.5 s sag to 0.70 pu. Diagnose the cause and describe the verification steps before re-testing.
Read the worked answer
Diagnosis. This is a settings problem, not a hardware limit. The agreement (derived from IEEE 1547-2018 and certified under UL 1741 SB) requires the inverter to stay connected through a 0.5 s sag to 0.70 pu. Tripping at 0.70 pu means the inverter is running a legacy trip curve or the manufacturer's default rather than the utility's ride-through profile.
Steps:
- Read back every protection setting from the inverter and compare with the utility's ride-through and trip table.
- Load the correct grid-support profile, then verify with a programmable AC source or a relay test set before reconnecting.
- Confirm anti-islanding remains enabled. Anti-islanding detects loss of the utility; a voltage sag is not an island, so it must not be disabled to pass the ride-through test.
- Re-test ride-through and anti-islanding, capture the plots, and submit them with the settings file to the utility.
Watch the firmware. A firmware update can silently revert protection settings. Lock the final settings and archive the file with the firmware version so the as-commissioned state is reproducible.
3. Evaluate the capacity test result and the SoC discrepancy against the vendor tolerance, and describe the acceptance action.
Read the worked answer
Capacity. 926 kWh ÷ 1,000 kWh = 92.6% of nameplate, inside the minus 5% beginning-of-life tolerance. The 2 h 4 min duration implies about 448 kW average, below the 500 kW rating - a lower C-rate usually yields more energy, so the result is conservative. Confirm the test started at 100% SoC, ended at the vendor's cut-off, and ran within the temperature limits of the agreed procedure; otherwise the number may not be contractually valid.
SoC drift. A 4% error after a full cycle points to a drifting coulomb counter or an algorithm that never resets at the top of charge. A 4% error is enough for the dispatch controller to hit the low-SoC limit early or over-discharge, so it must be fixed before commercial operation. Recalibrate with a full charge to 100% plus a hold, then a full discharge, and compare measured amp-hours with the BMS value.
Also record round-trip efficiency (AC out ÷ AC in) separately. DC capacity alone does not demonstrate round-trip performance, and the RTE number is what the warranty and the customer's economics depend on.