PV yield loss: systematically find and remedy causes

The solar installation on the commercial roof is producing less than planned – but nobody knows why. The inverter log looks normal, modules appear clean, monitoring shows no obvious errors. This is precisely the situation many operators find themselves in. Yield losses arise from a large variety of causes, most of which cannot be identified without specialised inspection methods. This guide explains the most common causes, how to diagnose them and – crucially – which inspection method is the right one for which symptom.

Normal degradation vs. technical defect – the first step

Not every yield loss is a defect. Solar modules lose a small portion of their power every year through physical ageing processes. This natural degradation is normal, predictable and priced into manufacturer guarantees. Typical values are 0.3–0.5% power loss per year. After 10 years this means a normal decline of approximately 3–5%.

If your system is producing 6% below original rated power after 8 years this is no cause for concern. If it has already lost 12% after 3 years there is very likely a technical defect or system fault.

The first step to diagnosis is therefore comparing the actual yield with a realistic projection. Tools such as PVGIS from the European Commission calculate the expected annual yield for every location in Europe free of charge. If your system permanently deviates more than 10% from this projection – taking into account annual irradiance fluctuations of ±5–8% – a cause analysis is appropriate.

The most common causes of yield loss

Yield losses in PV systems have very different causes. Some affect individual modules, others the entire string or the whole system. Understanding the cause categories helps in selecting the right diagnostic method.

Module level – individual modules or cells: Hotspots from cell degradation, microcracks from mechanical stress, delamination of the encapsulation film, snail trails (silver migration tracks), PID (potential-induced degradation) and soiling from bird droppings or industrial deposits. These defects usually affect individual modules or module groups and are only visible in the monitoring as a diffuse overall loss.

Stringebene – Verschaltungsfehler: Faulty diode circuitry in junction boxes, increased contact resistance at connectors, string mismatch from combining different modules or differently shaded modules in the same string. These errors affect entire strings simultaneously.

Anlagenebene – Systemfehler: Incorrectly calibrated or degraded inverter (particularly after 10+ years), suboptimal MPP tracking parameters, oversizing or undersizing of the DC side, line losses from cable cross-sections that are too small. These errors reduce the overall yield of the system.

Externe Einflüsse: Increasing shading from vegetation, neighbouring buildings or erected objects, changed reflection conditions, permanent soiling near industrial facilities. These causes are difficult to isolate with technical inspection methods and require comparative analysis over time.

Performance ratio: how to assess your system

The performance ratio (PR) is the most important metric for assessing a PV system. It relates the energy actually generated to the theoretically possible energy at the measured irradiance. A PR of 1.0 would be physically ideal; real systems achieve 0.75–0.88.

PR-Wert	Bewertung	Handlungsbedarf
≥ 0,82	Sehr gut	Routineinspektion planmäßig
0,75–0,82	Gut bis befriedigend	Investigate cause when trend deteriorates
0,68–0,75	Auffällig	Thermography and system inspection recommended
< 0,68	Critical	Sofortige Diagnose erforderlich

For the PR calculation you need: actual annual production in kWh, installed rated power in kWp and irradiance at the location in kWh/m² from the monitoring or from satellite data (PVGIS, Solargis). Your inverter or monitoring system should provide all these values.

Prüfmethoden im Vergleich

Several methods are available for diagnosing yield losses. No single method covers all causes – the choice depends on the symptoms and system size.

Monitoring-Auswertung is the first step and is free of charge. String monitoring shows which strings are anomalous. Without string monitoring the inverter analysis only provides total values. Diagnosis depth: system level. Limitations: no module level, no defect identification.

Drone thermography to IEC TS 62446-3 captures all heat-related defects at module level during live operation. Hotspots, microcracks, delamination, PID and faulty connectors are identified and classified. Diagnosis depth: module level and BOS level (in the Complete package). Limitations: no information about the electrical characteristics of individual modules.

Kennlinienmessung (I-V-Kurve) measures the electrical characteristics of individual strings or modules. Uncovers defects that leave no thermal fingerprint (certain forms of degradation, MPP deviations). Diagnosis depth: string and module level electrically. Limitations: laborious for large systems, requires shutdown of individual strings.

Elektrolumineszenz (EL) makes microcracks and cell defects visible at high resolution that are not detectable in the thermogram. Diagnosis depth: cell level. Limitations: system must be shut down, not drone-suitable for ground-mounted systems, high effort.

Entscheidungsbaum: Welche Prüfung wann?

What thermography detects for yield losses – and what it does not

Thermography is the first-choice method for diagnosing the causes of yield losses – but it has defined limitations. What it reliably detects: all defects that generate electrical resistance and thus produce heat. This covers hotspots of all causes (microcracks, cell defects, shading), delamination beyond a certain extent, PID patterns, faulty connectors and overheating in inverters and junction boxes.

What thermography does not detect: homogeneous degradation that affects all cells uniformly and produces no temperature difference, minor power drops below 3–5% without a thermally measurable origin, and defects that only occur under certain operating conditions (such as at high temperatures or specific irradiance angles). For these cases an I–V curve measurement or EL inspection is the right approach.

Economics of diagnosis

The question of whether a diagnosis is worthwhile can be answered with a simple calculation. Assume: a 200 kWp system produces 10% less than expected. At a feed-in tariff rate of 8 cents/kWh and 1,000 full-load hours this gives an annual yield shortfall of 200 kWp × 1,000 h × 0.10 × €0.08/kWh = €1,600 per year. Over 10 years this cumulates to €16,000 in lost income.

The costs of a complete thermographic inspection for this system size are in the range of €950–1,200 net. The ratio of diagnosis costs to lost income is clear – if the cause can be identified and remedied the inspection pays for itself in the first year.

Frequently asked questions