By John F. Jordon
Regardless of where you stand on issues such as trade disputes, feed-in-tariffs and government subsidies, one thing is certain: The drive for less expensive, more efficient modules is inexorable. Competition has increased in every way: Customers expect better modules at lower cost with stronger performance guarantees. In fact, these guarantees are becoming more impressive every year. Coverage extending for longer periods of time with no performance degradation is now the norm.
As the competition becomes stiffer, module manufacturers are challenged to deliver better designs with less expensive materials faster than ever. At a minimum, the profitability of their firm could rely on getting the next product release out on time; and the worst case, it could put the viability of the firm in question.
The proposition of stiffer competition with better products at a lower cost sounds like any mature business. Consumer electronics, housing, automobile manufacturing, even food and pharmaceuticals have the same challenges.
What¡¯s unique to the Photovoltaic (PV) module business is that the products being released have such a long operating life. Think about the consumer electronics industry, where accelerated testing relates to a one- or two-year warranty at best. As another example, a car is reasonably expected to last about 10 years and over 100,000 miles even though the first owner rarely has it that long.
A PV module is certainly much less complex than a car, computer or a cell phone. But PV modules are being guaranteed to operate with virtually no performance degradation for between 10 and 25 years¦¡which you can¡¯t say about a car, computer or cell phone.
Given that the average life span of all companies in developed nations is about 12.5 years, this is incredible. Worse¦¡the average life span of a company on the S&P 500 is only 25 years! This means that if you purchase a module from the most stable company today, you have only a moderate chance that they will be around in 24 years to honor the warranty.
Even though the PV industry has been around for 20 years now and there are over 50 GW of PV power-generation installed worldwide, it¡¯s still considered a ¡®nascent¡¯ technology. As suppliers and manufacturers within the industry, we often forget that the real competition isn¡¯t between the module manufacturers, installers or component suppliers. Sure, we all compete for any particular piece of business, but the real competition is in potential customers NOT installing a PV system, but choosing an alternate power generation technology instead.
So much time is spent discussing improved efficiencies, reduced system costs, steeper price curves and trade-offs in cost-outs that we sometimes forget another factor: To install a system, the customer must perceive it as being both cost competitive and reliable.
So even if every module manufacturer today did survive and were fully capable of replacing a failed module, the industry simply can¡¯t afford to have modules failing. It¡¯s really that simple. Failed modules don¡¯t just hurt the company that made them; they hurt the entire industry.
Most companies fully understand the math of failed modules, even if their only interest is in preserving their bottom line. In fact, one could argue they understand it so well they all agree on this point. The liability to any single manufacturer is simply too great not to put into place a robust component and finished-product testing program.
I¡¯ll even bet that some readers are thinking that all of this is obvious, so what¡¯s the point?
Here¡¯s the irony: if everyone agrees that failed modules are unacceptable, and accelerated testing is critical to the viability of the business, why aren¡¯t there standard tests that accurately predict PV module service life?
The answer may lie in how important accelerated testing is to the industry as a whole. Interestingly, many companies have taken the testing program to the point of being a competitive advantage. For these companies, their thinking goes something like this:
-If my testing is better than yours¡¦
-And if I can demonstrate through some testing that my product lasts longer than yours...
-Then I can say my product is better than yours¡¦
-Then I¡¯ll certainly sell more.
And then I win and you lose.
So instead of coming together to create test standards, companies are competing with each other on their testing.
Backsheet as a Component Example
It might be interesting to look at a single component of the solar module as an example of what¡¯s going on. DUNMORE makes backsheets, so that¡¯s what we¡¯ll use.
For those that don¡¯t know what a backsheet is, the name describes it pretty well. It¡¯s the back of the solar module, usually a laminate of different polymer materials. The backsheet acts as an electrical insulator and must also protect against the following impacts:
-Temperature swings from -40¡ÆC to +85¡ÆC
-Humidity and Vapor
-Dryness / Wind / Dust /Sand/Chemicals
-Scratches during installation and maintenance.
The backsheet laminate was originally made about 30 years ago using Tedlar¢ç, a fluorinated polymer developed and sold by DuPont. The original backsheet construction was Tedlar/Polyester/Tedlar (TPT). It turns out that this design has been proven to be totally robust by the test of time¦¡so robust, in fact, that many modules constructed with TPT backsheet from 30 years ago are still in service and going strong. So why not just make all backsheets a TPT construction?
Unfortunately, as the module business expanded, so did many other businesses that use Tedlar¦¡such as aircraft manufacturing. The demand for the material outstripped supply and a wide variety of alternative constructions have been developed. The list is actually quite long, including non-fluorinated constructions, alternative Tedlar constructions and fluorine-coated products.
Market requirements have driven these developments and many companies are using alternatives to TPT. But if TPT is proven for 30 years, how can we ensure that these alternate constructions will last?
Underwriters Laboratories and the International Electrotechnical Commission (IEC) have both developed a variety of tests to help the module manufacturers, and everyone in the supply chain has developed tests of their own.
Here¡¯s a sample list of some of the tests currently in use:
-Damp Heat Test (typically 85¡ÆC and 85% r.h.)
-Partial Discharge Test
-Peel Test (typically at 90¡Æ & 180¡Æ)
-UV-Test (IEC 61215)
-Shrinkage Test (both temperature and time dependant)
-Water Vapor Barrier Test
-UV-Damp Heat Combination Test
Can We Accelerate the Accelerated Test?
Now, let¡¯s just take one example and drill into it a little bit. It is well proven that testing at 85¨¬C and 85% relative humidity for 1,000 hours (about 42 days) will reveal potential failure modes in backsheet laminates. IEC Standards 61215 and 61646 define the parameters, procedures and criteria for this testing.
Almost every backsheet on the market today can pass that test. In fact, it¡¯s now considered a relatively low bar. What¡¯s missing is a definition of test standards for polymeric components. For example, bond test procedures and the criteria specific to polymeric backsheet components. As a result, module manufacturers now routinely test at 85¨¬C and 85% R.H. to 2,000 hours and even 3,000 hours.
Wait a minute, 3,000 hours is 125 days. A test that used to take 42 days is now more than four months. The financial guys are shouting ¡°just when we need to take cost out, we¡¯re tripling the test time for a material! This is going in the wrong direction! Why not just test it for 30 years in the field?¡±
Now, there are a lot of bright people in the PV industry, so some research has been done to show that you can put the backsheet in a pressure cooker for several days and induce similar failure modes to those found in 3,000 hours heat and humidity. Meanwhile, others have observed that you can put backsheet into boiling water for a few hours and also induce similar failure modes.
So we¡¯ve actually gone from a 42-day UL test to boiling in water for a period of minutes? Can this possibly work? Can we even wait for the water to boil?
Unfortunately, there is limited data to show the correlation between these tests and their failure modes. There is even less data to suggest these tests adequately predict module life.
Gregg K. Hobbs, who wrote the ¡®Accelerated Reliability Engineering, HALT & HASS¡¯ (Wiley copyright 2000), and is considered to have developed key tenets of accelerated testing, has defined the problem. He clearly says that accelerated testing is meant to identify failure modes¦¡not to be used as a means of qualifying materials. He says, ¡°The stresses applied in HALT and HASS are not meant to simulate the field environments at all, but are meant to expose the weak links in the design and processes using only a few units and in a very short period of time.¡± (Hobbs, p. 5).
He also discusses the difference between Mean Time Between Failure (MTBF) testing and Highly Accelerated Lifetime Testing (HALT). Some engineers and scientists are observing that the PV industry seems to be caught in the middle: we don¡¯t have the data to fully support MTBF testing, so we¡¯re supplementing it with HALT testing.
More importantly¦¡what if this testing doesn¡¯t work? What if in the rush to show that your product is ¡®better¡¯ than the competitors, you are simply undermining the entire industry?
It might not just be one company that suffers, but the entire solar module market.
Collaboration Is the Answer
Let¡¯s switch gears for a minute and talk about those other industries. Dunmore Corporation is a broad-based company that has been in business for over 40 years. Not only do we make backsheet for solar modules, but we also supply other products that require a long life¦¡such as films for airplanes, the Space Program, and products that go into cars, computers and cell phones.
These more mature markets that we¡¯re engaged in have well-defined UL, TUV or ANSI standards. In the case where these standards don¡¯t apply, a government body such as NASA or the FAA has a test standard that is clear.
Other industries almost universally include accelerated testing or simple failure tests that are well proven and relate to the expected life of the product. More importantly, we don¡¯t compete¦¡and our customers don¡¯t compete¦¡on these test procedures. In fact, it¡¯s not even a discussion. It¡¯s in the product standard, it¡¯s part of our quality system and we get audited on our performance. It¡¯s a price of entry¦¡not a means of differentiation.
So why can¡¯t we do that for a solar module? The short answer is we can, and it¡¯s already happening.
A group of material suppliers and module manufacturers have already gotten together with the goal of minimizing national differences by characterizing testing for safety and reliability.
This particular group actually chose a pretty appropriate name; it¡¯s called the ¡®International PV Module QA Task Force¡¯. They are comprised of over 200 volunteers from throughout the industry, working to define tests that can be used to predict module service life. The membership overlaps between UL, IEC and WG2 modules groups ¦¡among others.
When this group has completed their work, all PV component data sheets will require global standards test results to mitigate risks of module manufacturers and the solar industry.
These individuals and the companies they represent deserve a great deal of credit¦¡and they also deserve your support. This should be a model for the rest of the industry¦¡not just for those directly impacted by this group.
As a member of the PV community, regardless of your position in the supply chain, I would ask you to consider taking the following three actions:
1. The next time your company wants to sell (or buy something) based on the result of a singular element of test data please rethink it. We should all be selling a quality product¦¡built to last for up to 30 years, but the offer should be based on a broad range of standardized tests that support all critical product requirements.
2. Start moving toward making material decisions on MTBF data, not just on HALT or HASS data. Accelerated testing clearly has its place¦¡but if the failure mode only occurs in the accelerated test and not in the field, the test is not useful.
3. Take a look at what the International Module QA Task Force has done and see if it can be applied to other sectors of the PV market. By cooperating in a similar way to other industry groups we can arrive at accepted standards for the greatest longevity of our products¦¡and, ultimately, dramatic growth of the solar market.
John F. Jordon joined DUNMORE in 2010. Previously, he was President, worldwide field operations, for TAGSYS, a leader in RFID tagging solutions. He also initiated a ¡®Quality of Service Program¡¯, assuring that client installations achieved six-sigma performance levels. Prior to this, Jordon was Vice President of international operations for Apriso. In 1997, he sold and implemented the first RFID system in the U.S. for Gemplus Tag (now Gemalto). Before the RFID market, he designed and operated a plant for BOC Gases and oversaw engineering, operations, distribution and supply chain management. Jordan earned a master¡¯s degree in Chemical Engineering from Rutgers University.
For more information, please send your e-mails to firstname.lastname@example.org.
¨Ï2011 www.interpv.net All rights reserved.