BRENNEN

Industries / Energy, water and waste / Power generation / Solar panels

Testing and reliability

Testing and reliability checks that a panel delivers the power on its label and keeps working for 25 years or more, using lab stress tests, field records and independent scorecards.

7

Milestones achieved, from 1981 to 2026.

1

To watch for: announced, not yet confirmed outside.

2

Open deliverables. Nobody has shipped them.

The road

AchievedWatch forOpen

Achieved

  1. 1981

    Tougher tests cut failures

    From 1976 NASA's Jet Propulsion Laboratory (JPL), working for the US Department of Energy, bought modules from industry in five rounds called block buys; each round set tests a design had to pass before its maker could deliver in quantity. The first tests were borrowed from space arrays, such as heating and cooling between −40 °C and +90 °C. Block V, from 1981 to 1985, raised these thermal cycles from 50 to 200 and added a humidity-freeze test (hours at 85 °C and 85% humidity, then −40 °C) and a hot-spot test; a hail test with 25.4 mm ice balls had come with Block IV. Field records reported in 1993 showed a 45% failure rate for modules designed before Block V and under 0.1% for those after, according to John Wohlgemuth of NREL, who says JPL's unfinished Block VI became the basis of IEC 61215, the international qualification test.

    Lessona short, strict set of stress tests removed most early failures, but by design it does not measure how a module wears out over decades.

    ntrs.nasa.govosti.gov

    Credit

    JPL's Flat-Plate Solar Array Project; Whipple (1993 field review)

  2. 2010

    High voltage drains cells

    As arrays were wired to higher voltages, current leaking from the cells through the glass to the grounded frame cut their power. Pingel and colleagues at SOLON described it in a 2010 paper titled "Potential Induced Degradation of Solar Cells and Panels", and the name potential-induced degradation (PID) stuck. NREL's Peter Hacke and colleagues wrote in 2011 that JPL had listed voltage as a stress to test for in 1978, but that the qualification standards left such a test out as too harsh, an omission that "has cost the industry significant trouble and expense." NREL then led a round-robin test among labs for a draft standard, IEC 62804, that holds modules at system voltage in 60 °C and 85% humidity for 96 hours.

    Lessontests designed for small low-voltage systems missed a failure that only large high-voltage arrays caused.

    osti.govosti.gov

    Credit

    S. Pingel and colleagues, SOLON; Peter Hacke and colleagues, NREL

  3. 2012

    Half a percent a year

    Dirk Jordan and Sarah Kurtz of NREL gathered nearly 2,000 degradation rates (how fast a module's output falls) measured in the field over 40 years and found a median of 0.5% a year. Their 2016 update, with more than 11,000 rates from almost 200 studies in 40 countries, put crystalline silicon modules at a median of 0.5 to 0.6% a year and a mean of 0.8 to 0.9%.

    osti.govosti.gov

    Credit

    Dirk Jordan and Sarah Kurtz, NREL

  4. 2012

    Light and heat sap cells

    K. Ramspeck and colleagues reported in 2012 a new power loss in cells with a passivated rear surface, later named light and elevated temperature induced degradation (LeTID). In 2020 NREL traced the loss at a utility plant in the US Mid-Atlantic, where five of six arrays of the same module model had degraded faster than the sixth over about three years, to LeTID, after field measurements on 224 modules and lab tests on a few of them.

    Lessona cell change made for efficiency can bring a loss that shows up only after years outdoors.

    osti.gov

    Credit

    K. Ramspeck and colleagues (first report); Michael Deceglie and colleagues, NREL, with Core Energy Works and Constellation (field case)

  5. 2020

    Backsheets crack outdoors

    Cracking of polyamide backsheets (the plastic rear layer of glass-backsheet modules) was widely reported in modules in the field, in places including China and Italy, although standard accelerated tests had not caught it. NREL and SLAC researchers reproduced the cracks only with tests that applied several stresses together or in sequence. By 2026 Kiwa PVEL had recorded no catastrophic backsheet failures in its own backsheet tests and called this an example of the industry "seemingly solving a reliability issue."

    Lessontests that apply one stress at a time can miss failures caused by sun, heat and moisture acting together.

    osti.govscorecard.pvel.com

    Credit

    Michael Owen-Bellini and colleagues, NREL and SLAC; Kiwa PVEL

  6. 2024

    Labels promise too much

    Fraunhofer ISE's calibration lab in Freiburg compared 6,349 of its power measurements of crystalline silicon modules since 2012 with the power on their labels. The average shortfall grew from 0.4% in 2016 to 2.2% in 2023, and modules measuring above their label had nearly disappeared. For modules from large makers, it reported a shortfall of 1.3% in 2023 and 1.2% in 2024.

    pv-magazine-australia.compv-tech.org

    Credit

    Daniel Philipp, Martin Kaiser and Ulli Kräling, Fraunhofer ISE

  7. 2026

    Most makers fail something

    Kiwa PVEL, which has run its module testing program since 2010, reported in November 2026 that 87% of manufacturers and 61% of module recipes it tested had at least one failure, up from 83% and 59% a year earlier. Modules breaking in its mechanical-load and hail tests was the most common failure, and the share of manufacturers with a delamination failure (layers coming apart) tripled to 45%. Ultraviolet light still damages some newer TOPCon cells, a loss called UVID: the median loss for TOPCon modules in Kiwa PVEL's ultraviolet test fell from 4.2% to 3.0%, and some modules left outdoors in Davis, California, lost up to 8% in their first year.

    scorecard.pvel.comscorecard.pvel.com

    Credit

    Kiwa PVEL

Watch for

  1. A test for ultraviolet damage

    An International Electrotechnical Commission (IEC) group is writing IEC TS 63624-1, a technical specification for lab tests of ultraviolet-induced degradation in silicon modules, including procedures for the part of the loss that recovers in light. Trina Solar researchers have proposed splitting the loss into reversible and irreversible parts to match it. Kiwa PVEL says its own ultraviolet test already goes beyond what IEC and UL certification requires.

    Achieved whenthe IEC publishes the specification and an accredited lab offers testing to it.

    pv-magazine.comscorecard.pvel.com

    Claimed by

    Trina Solar researchers, reported by pv magazine, October 2026

Open

  1. 01

    Check the certificate claims

    A software agent that reads a module datasheet, lists each safety and quality certificate it claims (such as IEC 61215, IEC 61730 or UL 61730) with its number, looks each one up in the issuing body's public register, and checks that the certificate exists, is current, and covers that model, power range and factory. It publishes a tally.

    The brief
    Credit · unclaimed
  2. 02

    Bench-test junction boxes by name

    A published bench test of the junction boxes sold to module makers (the sealed box on the back of a module that holds the bypass diodes, which let current skip a shaded or broken string of cells). Buy at least 10 models, run current through their diodes while cycling the temperature, log diode temperature and voltage, and take each box apart afterwards. All at low voltage, on boxes not attached to any panel.

    The brief
    Credit · unclaimed

Open 01 · the brief

Check the certificate claims

Credit · unclaimed

The first to publish is credited on this line, and the deliverable moves from Open to Achieved.

What to build

A software agent that reads a module datasheet, lists each safety and quality certificate it claims (such as IEC 61215, IEC 61730 or UL 61730) with its number, looks each one up in the issuing body's public register, and checks that the certificate exists, is current, and covers that model, power range and factory. It publishes a tally.

Why it is still open

In 2015 TÜV SÜD told PV Tech that a QSolar module carried a counterfeit TÜV SÜD mark for a model it had never certified. In 2024 European customs seized fake Bauer modules that had no certification. The California Energy Commission says makers' self-reported information on its solar equipment lists "is not confirmed." Australia goes furthest: under its Solar Panel Validation initiative, installers scan each panel's serial number against makers' databases to confirm it is an approved panel, but only for small systems claiming the national rooftop subsidy. Registers such as TÜV Rheinland's Certipedia can be searched one certificate at a time; a search found guidance for checking by hand but no tool or published tally that checks in bulk.

pv-tech.orgpveurope.euenergy.ca.govcer.gov.aucertipedia.com

Built with what exists

A browser agent (see the Note "Deleting 80% to start over"); a language model that reads documents (see the Note "Seven more flagged"); Python; the certifiers' public registers; a person who reviews every mismatch.

  1. First step

    Take 20 datasheets from one distributor's catalog, check their certificates by hand on the registers, and time each lookup.

  2. Done looks like

    500 datasheets checked, with every mismatch confirmed by a person and sent to the maker and the certifier before anything is published; a tally of certificates found and matching, found but not matching, and not found; and an outside person who reruns 50 checks and gets the same answers.

Open 02 · the brief

Bench-test junction boxes by name

Credit · unclaimed

The first to publish is credited on this line, and the deliverable moves from Open to Achieved.

What to build

A published bench test of the junction boxes sold to module makers (the sealed box on the back of a module that holds the bypass diodes, which let current skip a shaded or broken string of cells). Buy at least 10 models, run current through their diodes while cycling the temperature, log diode temperature and voltage, and take each box apart afterwards. All at low voltage, on boxes not attached to any panel.

Why it is still open

Kiwa PVEL's 2026 scorecard says 17% of manufacturers had a failed bypass diode in its tests, 9% had a safety failure traced to the junction box, and that it has received more and more reports of junction boxes overheating or melting in the field, some leading to vegetation fires; it gives only totals, not models. A 2013 study by NREL and Trina Solar found that the diodes in one of three unnamed boxes failed outright under temperature cycling with current, and warned that such diodes may pass IEC 61215. A search found no published results by junction box model.

scorecard.pvel.comdigital.library.unt.edu

Built with what exists

A programmable low-voltage DC power supply rated for about 20 A; thermocouples and a thermal camera; a small temperature chamber; a Raspberry Pi data logger; the bypass-diode thermal test in IEC 61215; a microscope for teardown.

  1. First step

    Buy three boxes of one model, run rated current through one diode in a 75 °C oven for one hour, and log its temperature.

  2. Done looks like

    10 named models, three samples each, through the same sequence, with diode temperatures, changes in diode voltage, failures and teardown photos published, and each supplier shown its results 30 days before publication.