8 kWp Array Delivering 2.9 kW While the DC Side Looks Healthy
An 8 kWp grid-tied system was commissioned the previous day: two strings of eight 500 Wp modules (Voc 50.2 V, Vmp 42.0 V, Isc 12.6 A, Pmax temperature coefficient −0.35 %/°C, Voc temperature coefficient −0.29 %/°C) on a south-facing roof with no shading. On a clear June afternoon the irradiance is 850 W/m², ambient temperature 38 °C and cell temperature about 65 °C. The inverter app shows 2.9 kW AC while a neighbour's 8 kWp array is producing 6.1 kW. Field measurements: string 1 Voc 356 V, string 2 Voc 354 V; string 1 Isc 10.1 A, string 2 Isc 10.0 A. The inverter log holds twelve 'Grid voltage high' warnings from the past 24 hours. AC voltage at the inverter terminals is 252 V while the DISCOM meter at the service entrance reads 238 V; the AC cable is 4 mm² copper and the run to the service entrance is 35 m. The customer is refusing to sign the commissioning certificate.
1. Use temperature-corrected expectations to decide whether the DC side is responsible for the shortfall.
Read the worked answer
Expected Voc at 65 °C cell Per module: 50.2 × (1 − 0.0029 × 40) = 50.2 × 0.884 = 44.4 V; per string of eight = 355 V. Measured 356 V and 354 V - a perfect match. The strings have the right number of healthy modules, correct polarity and no open bypass path.
Expected Isc at 850 W/m² 12.6 × 0.85 = 10.7 A, plus a small positive temperature effect on current → about 10.8 A. Measured 10.0–10.1 A, i.e. ~94 % of expectation, consistent with light soiling on a June roof in Punjab. Nothing here justifies a 50 % loss.
Expected power 8 kWp × 0.85 (irradiance) = 6.8 kW; temperature derate = 1 − 0.0035 × 40 = 0.86 → 5.85 kW DC, about 5.5 kW AC after inverter losses. The neighbour's 6.1 kW is in the same band.
Conclusion: the DC side is exonerated. Voc and Isc both land where physics says they should, so the fault is downstream on the AC side or in the inverter's grid-interaction behaviour. Do not start swapping modules or diodes - that is the expensive way to learn nothing.
2. Explain the mechanism behind the 2.9 kW output using the AC voltage evidence, quantify it, and state the remedies in order of priority.
Read the worked answer
The mechanism. A grid-tied inverter pushes current into the grid until its terminal voltage reaches the upper limit (typically 253 V). If the impedance between inverter and meter is high, a small export current produces a large voltage rise, the inverter hits the limit early, folds back and settles at whatever power keeps it just under 253 V. The twelve 'Grid voltage high' warnings are the fingerprint; the array is fine but is being throttled.
Quantify. 252 V − 238 V = 14 V at roughly 2.9 kW / 230 V ≈ 12.6 A, implying a loop resistance of about 1.1 Ω. 70 m of 4 mm² copper should be 0.0172 × 70 / 4 = 0.30 Ω, i.e. ~3.8 V at 12.6 A. So roughly 0.8 Ω - about 10 V - is in joints, not in the cable: loose ACDB or MCB lugs, a badly made termination, or a corroded meter-tail connection.
Remedies in order
- Isolate, torque every AC terminal to the manufacturer's value, remake any suspect lug; re-measure the difference.
- Replace 4 mm² with 10 mm² (0.120 Ω → ~1.5 V at 12.6 A) - or convert to a three-phase inverter, which is the correct fix at 8 kW.
- Ask the DISCOM to check the distribution transformer tap and the feeder voltage; 238 V at no export is already high on a rural feeder.
- Only as a stopgap, set a temporary export limit and record it - never raise the inverter's over-voltage trip setting to hide the problem.
3. List the checks and records you must complete before you sign this commissioning certificate, including the safety steps for the AC work.
Read the worked answer
Electrical tests
- Insulation resistance of each string to earth at 1000 V DC (expect ≥ 1 MΩ, normally tens of MΩ) with the SPD and inverter disconnected.
- Polarity and continuity of every string; earth continuity from module frame to structure to earth pit, and earth pit resistance (target under 5 Ω).
- Per-string Voc and Isc at measured irradiance and cell temperature, converted to STC for the record - this is what proves the DC side is sound.
- Functional checks: anti-islanding trip, DC isolator operation, SPD status indicators, inverter firmware and grid-code setting.
- Thermal scan of the DC connectors, fuses and AC terminals under full load.
Safety for the AC re-termination
- Open the AC breaker at the ACDB, lock and tag it, and confirm dead at the terminals with a proving unit or a known-live source first, then the meter.
- Shut the DC isolator before the AC work so the inverter is fully off; wait for the DC-link capacitors to discharge before opening the inverter cover.
- Insulated tools, gloves and arc-rated clothing; one person works, another watches, and nobody works alone on a live consumer unit.
Handover records: test sheet with instrument serial numbers and calibration status, thermal images, single-line diagram, module and inverter serial list, net-meter sealing certificate, and a signed snag list. The customer should sign only after seeing the corrected power at the meter.