Success stories & chamber environments

Observations at a QA conference

David Jung, Marketing Manager, ESPEC North America, December 2011

Conference speaker Mr. Tadanori Tanahashi, Development Division, ESPEC

The emerging solar panel (photovoltaic) industry is under extreme pressure to be able to prove long-term quality of their products, as the panels are justified and financed based on usable lifetimes of twenty years or more. Up to now, a reliance on IEC and UL standards to weed out inferior PV panels has been successful. However, the consensus is that these standards are not sufficient to guarantee long-term quality.

US and Japanese government research arms NREL and AIST (respectively), struck with the need for universal and useful quality assurance standards, have created the “International PV Module Quality Assurance Forum”. The QA Forum is a true crowd-sourcing project, with a combination of live forums, online meetings, and wiki documents. Participation is altruistically motivated for the good of QA professionals and the industry as a whole.

Environmental testing, and temperature/humidity specifically, quickly has become an important focus of this group. How should the existing test requirements of temperature cycling, humidity freeze, and damp heat be improved?

ESPEC has been supplying test chambers for these tests, and certainly is interested in any changes that may come. Also, our company also has staff doing research into different test solutions for solar cells that may be useful for the industry. For those reasons, we have been participating in the QA Forums and I recently attended the second of the major meetings, held in Tokyo, Japan.

While there is much I could write about the technical issues themselves, I am no expert, and may not apply to the reader. However, it may also be instructive to record some observations in general. What would a group like this, meeting together, talk about, or set out to do?

We need to speak the same language
Amazingly, the first and most important question is this: “What is quality?” There are several parts of quality, and each has its own unique tangents. I think the group would agree that all aspects need to be addressed, but this can become confusing. The confusion arises because the same testing solutions can be used to address the different issues of quality.

Kazuhiko Kato, from AIST, specifically made the point that ‘long term reliability should be defined as safety’. Solar panels generate energy, and potential failures as they age include overheating and spontaneous fires. This type of failure is not an issue with most consumer products, or even industrial products, hence the overt warning that reliability should be defined as safety.

The present danger is that no one will care about safety … until it is a problem. Being in Japan, it was easy for the speaker to make the case for safety with one word: “Fukushima”.

Besides safety, reliability means sustaining the required function. PV panels generally degrade in performance as they age, and losses of performance for installed systems are caused by failures of sub-components or panels. To that end, Kato made the suggestion that panels be measured (or judged) by their potential output over their lifetime.

In an attempt to bring these issues together, Kato presented a new acronym to define quality for PV panels:

PVRessQ = PV Reliable Safe Sustainable Quality
In essence, to save everyone from the risk of using a generic word (quality) with multiple meanings, he created a completely new one. For such a group project as the QA Forum, this may help accelerate the discussion if the terminology gains acceptance.

Tony Sample, from EU DG-JRC, hit upon another aspect of quality: ‘long term quality relies on design, manufacture, installation, and environment’. Obvious, yes, but in the case of the QA Forum, worthy of making sure that all facets are considered. This is especially true because the current UL and IEC testing and certification methodology focuses on design. Installation may be the bastard-child of the four factors: uncontrolled, and highly variable … and the hardest to test for!

Choosing a new test, not so easy
As the focus of the Forum became more technical, presentations about testing to demonstrate long-term reliability came with a variety of potential solutions:
  1. One speaker was already doing “double IEC”, doubling the testing periods of the IEC standards.
  2. Another suggested ‘sequential testing’, completing the environmental tests on the same samples.
  3. The significance of UV light exposure combined with damp heat (85/85) was shown.
  4. Suggestions were made for environment-specific tests like wind or snow loading, or regional-specific qualification.

Tony Tang, from Suntech, said this about any new test standards: ‘Make them REASONABLE, PLEASE’. All-caps are in my notes, so it was either that way on his slides or in the tone of his speech. His point is part of the problem the QA Forum is facing, any new testing will need to be more extreme (and thus unreasonable) in some way in order to get meaningful results.

A presenter said that their experience (across many years, and brands) is that the older PV modules are less likely to fail … they are of higher quality. Of course, they were much less efficient, and much more expensive. It may be an easy assumption to make that they were over-designed and carefully manufactured. But still, the question (that came from the audience) remains: ‘What made them better?’

The two blockbuster presentations
To an outsider like me (and as a marketer), I found a qualitative and quantitative review of customer experience very informative. While other presenters were trying to figure out how to win quality in the lab, Kato was learning what quality looks like in the field. His group did residential surveys and inspections of PV systems. One in seven users had to replace a panel within ten years, while others had dead panels they didn’t know about. Such inspection is hard to manage with residential customers, yet critical in learning quality.

From the opposite end, Dr. Xiaohong Gu, from NIST, showed results of her efforts to link accelerated lab testing and modeling (with UV and temp/humid) with real-world effects. Although her demonstration was only over short time periods, the ability to ‘predict’ the performance of solar panels was quite exciting.

Side-note on choosing testing conditions:
When talking to Mike Kempe of NREL before the event, he made a very strong point to me that when choosing to do additional testing choose DIFFERENT conditions, NOT more of the same. Why double the IEC test periods, in other words? Probably very true for a researcher, but I wonder if this philosophy will be useful for an industry under pressure to deliver results.

For any comment or inquiry regarding Test Navi