U.S. Department of Energy - Energy Efficiency and Renewable Energy

SunShot Initiative

Photovoltaic Module Reliability Workshop 2012

February 28–March 1, 2012

The Photovoltaic (PV) Module Reliability Workshop was held in Golden, Colorado, on Feb. 28–March 1, 2012. The objective was to share information to improve PV module reliability because such improvements reduce the cost of solar electricity and give investors confidence in the technology. NREL led the workshop, which was sponsored by the U.S. Department of Energy (DOE) Solar Energy Technologies Program (Solar Program).

The following documents are available. The entire proceeding of the PV Module Reliability Workshop (Files) will be available for download soon. In addition, a detailed agenda is available.

The presentations are grouped by topic and session:

Overview Presentations

Back to Top

Plenary Session: Silicon

Safety Issues

Chairs/Discussion Leaders: Jay Johnson (Sandia) and Chris Flueckiger (UL)

Discussion summary (safety issues):

  • The challenge is to construct a rooftop PV system that simultaneously keeps the panels cool and achieves a Class A fire rating.
  • When ANSI adopts IEC 61730, harmonizing the U.S. and International safety standards, national differences will be retained to include spread-of-flame tests of PV systems mounted on roofs.
  • Addressing series arc faults is useful, but will increase the issues with parallel arc faults. We should address both, and the technology for addressing the parallel arcs may be available soon.
  • Solar ABCs will work to develop a standard for inspecting older PV systems this year.
  • Suggestion for DOE to support a program to inspect old PV systems through SunShot and the RTCs.

Reliability Predictions through Analytical Modeling

Chairs/Discussion Leaders: Glenn Alers (UC Santa Cruz), David Meakin (Fraunhofer CSE)

Discussion summary (modeling):

  • Modeling far into the future tends to exaggerate testing anomalies and uncertainties.
  • Modeling helps us:
    • Improve product design
    • Explore our understanding
    • Extrapolate to longer times
  • We still need field experiments for:
    • Gathering statistics for manufacturing variations
    • Understanding combinations of factors
  • Stochastic approaches allow modeling of complex systems with less computational time.
  • We often take diodes for granted. We should explore this as an industry, especially considering the 25% failure rate seen in the Japanese study.

Potential Induced and Other Bias-Related Degradation

Chairs/Discussion Leaders: Mike Kempe (NREL), Jenya Meydbray (PV Evolution Labs)

Discussion summary (PID):

  • PID is observed in both mono- and multi-crystalline systems.
  • Could we develop a test for encapsulants that identifies whether they will help the mitigation of PID?

Regional meeting of PV QA Task Group #1

Back to Top

Poster Session: Silicon

Back to Top

Plenary Session: Standards

Materials Testing

Chairs: Kurt Scott (Atlas), Michael Koehl (Fraunhofer ISE)

Quality Assurance Overviews of Proposed Tests

Proposed Test Protocols — IEC 61215 on Steroids

Proposed Test Protocol — Accelerated Simulation of the Weather

Proposed Test Protocols — New Tests

PV Quality Assurance Task Force Plans

Quality Assurance Rating Test Development Discussion

  • Summary of Proposed Test Methods (see handout), Ian Aeby (Emcore)
  • Consensus Building — Can we agree on general items? What we can learn from proposed tests? Can we agree on detailed inputs for the Task Groups?

Discussion summary (QA standards):

  • The combination of stresses is most important. Any test we design should include combinations of stresses.
  • Testing concerns include:
    • The methods in which the tests are applied can yield different results.
    • Tests will take too long.
    • Tests may not reflect actual failure mechanism.
    • Tests may require frequent, expensive retesting.
    • New guideline could be confused with IEC 61215.
    • Rankings will not be related to product value in a way that customers can understand.
    • With faster temperature cycling, modules in the middle of the chamber would not be stressed as much as those close to the thermal control.
    • It is difficult to define a standard method for making a test coupon for UV testing.
    • UV testing for full-size modules is expensive.
    • UV testing could be accelerated too much, changing the failure mode.
    • Testing to IEC 61215 and to a new standard will increase costs.
  • Suggestions include:
    • Use a triangle or a circle and colors to communicate rating.
    • Design the highest rating to be for the most extreme stresses found on the earth.
    • Modify IEC 61215 instead of creating a new document.
    • Only the range of temperature change will affect the acceleration factor for Cu ribbon, while solder joints will additionally be influenced by ramp rate and dwell time.
    • Combine UV and mechanical testing.
    • UV test both full-size module (to test interconnects) and test coupons (to screen new materials).
    • Use weatherometer testing to identify yellowing problems.
    • Create a task force to collect field data.
    • Require every module manufacturer to bring data to the workshop next year.
    • Find other ways to require companies to share failure data.
    • Consider testing mechanical load at low temperature.
    • Consider illuminated stress testing of minimodules.
  • Participants agreed that:
    • Comparative testing is preferred over pass/fail testing because durability can be assessed for different climates. For example: 5 stars is better than 4 stars is better than 3 stars, etc.
    • Comparative testing is the first step to predicting module lifetime for a specific application or location.
    • Testing results should be communicated both in a short summary that can be recorded on the module and detailed data that is communicated in a report.
    • We need an option for subjecting the module to greater thermal-cycling stress than IEC 61215 (TG2).
    • We want to do more UV testing in test coupons that represent full package
    • The ranking must to be tied to field performance.
    • The new test standard should use a different number rather than be considered Part 2 to IEC 61215.
    • We need to apply mechanical stress before thermal cycling.
  • The majority expressed a preference for a rating system that describes module durability for a list of stresses (temperature, humidity, UV); some preferred describing the durability with respect to climates (hot and dry, hot and humid, etc.).
  • A key question is whether ultimately all of the manufacturers will make their products achieve the highest rating in the system. Cases in which a lower rating could enable lower costs:
    • Reduced thickness of glass when mechanical (snow/wind) load is not an issue
    • Low voltage systems, which use fewer bypass diodes
    • Locations requiring special (expensive) glass to avoid abrasion from the sand.
  • Manufacturers do not determine where the modules are located, so it may be best to have one design that works everywhere.
  • We need field data to better understand what tests are needed. If PV owners can share data confidentially, this may help guide the designs of the tests. For example, are we seeing failures related to humidity?
  • PID is observed less often in the U.S. because of the grounding. The new proposed test would allow modules to fail the test, but then be able to pass it simply by modifying the data sheet to require the system to be grounded.

Back to Top

Oral Session: Thin-Film Modules

Metastabilities

Chairs/Discussion Leaders: Chris Deline (NREL), Govindasamy Tamizhmani (TUV)

Keeping the Moisture Out

Chairs/Discussion Leaders: Dennis Coyle (GE), Arrelaine Dameron (NREL)

Discussion summary (thin films):

  • We need appropriate measurement techniques for metastable thin-film modules. We should retain the stabilization procedure in IEC 61646 as a general thin-film test procedure. Testing labs need to understand the limitations of the devices that they are testing, and, for example, measure the CIGS modules very quickly after the light soaking stabilization, as they have a fast relaxation time.
  • Responding to a statement at the DOE Metrology workshop, participants discussed whether the topic of metastabilities is still a priority, considering that the best CIGS modules show only very small metastabilities. Participants agreed that metastabilities are still a vital topic, but generalizing may be difficult due to the dependence on certain processing methods and conditions. For example, each product may respond differently to light and dark exposure.
  • We can increase stability by:
    • Annealing CdS after deposition
    • Increasing CdS thickness
  • Sometimes, higher efficiencies show smaller metastability.
  • We could question the relevance of the 85/85 damp heat test, but data is not available to define the necessary test.

Reliability Issues with Thin-Film Modules

Chairs/Discussion Leaders: Shubhra Bansal (DOE), Tony Sample (JRC)

Back to Top

Poster Session: Thin-Film Modules

Back to Top

Oral Session: CPV

CPV Module Reliability — Accelerated Testing and Field Experience

Chairs/Discussion Leaders: Peter Hebert (Spectrolab), Paul Lamarche (Solar Junction)

Standards

Chairs/Discussion Leaders: Kenny Villegas (Solar Compliance), Robert MacDonald (Skyline Solar)

Modeling of CPV Reliability Issues

Chairs/Discussion Leaders: Nick Bosco (NREL), Scott Burroughs (Semprius)

Discussion summary:

  • Testing at high fluxes is difficult. Will irregularities cause problems?
  • Soiling and condensation cause problems in the modules, but are not addressed by IEC 62108. IEC 60068-2-68 has defined a vertical dust flow test.
  • In the short term, trackers and inverters have more issues than modules, but module failure could be devastating in the long term.
  • The tracker standard attempts to address tracker issues, but more data is needed to know whether the draft standard will be effective.
  • Bypass diode testing should be defined as part of the cell-on-carrier test standard.
  • We should test whether the lenses lose alignment at low temperatures. This test might be done as part of the energy rating.
  • Most manufacturers do more testing than modeling because they don't have confidence that the modeling will give the correct result and believe that testing is a faster way to determine whether a design will work.
  • Modeling can reveal ways to improve a design. For example, modeling shows that matching the thermal expansion coefficients can help to reduce damage accumulation in thermal fatigue.
  • Company representatives value having the universities and national labs do modeling, allowing the companies to focus on testing their product, while gaining increased understanding from the modeling.
  • Modeling may help us understand whether chambers packed with modules can ramp temperatures uniformly and quickly.

Back to Top

Poster Session: CPV

Back to Top