A recent report (PDF) reminded us that the solar power industry is still in its youth. While we have been installing solar power since the 1950’s – 85% of the world’s 230GW+ has been installed in the last 5 years, and during the next five years that front end concentration will increase. The suite of tests developed in the 1970s for measuring solar module longevity is showing us that while most modules perform well, at or above their warranted levels, there are enough that don’t, and that the industry need continue to improve quality control if we are to depend on solar power in the long term.
When a large scale solar power plant is built out, the financiers of those solar plants are going to ask if your solar modules are ‘bankable.’ Bankability generally follows a few guidelines, the first being – What are the odds that the solar module will perform like the production models say? $100 billion of solar power was installed last year – and every installation believes their solar modules are going to last long enough to cover their costs.
Solar module failure occurs in three phases – Infant, Midlife and Wear Out. Generally, the Infant phase failures are found quickly during project installation. Most of the loss that occurs in Midlife is due to glass discoloring or cell delamination. Once into the Wear Out phase the connection between individual cells will corrode. During the entirety of their lifetime solar modules will be losing efficiency, degradation, at a rate of 0.2-0.7%/year.
In a field study done by NREL, the average rate of annual degradation was below 1.0%, however, 15% of solar modules had a rate of degradation above 1% a year (deemed a failure per the warranty). On the positive side – when removing that 15% of modules, the industry average was well below a degradation loss of 0.5%/year – a value at least 20% better than the generally warranted production. A full 10% of modules were still producing electricity at or above their original warranted level at varying years in the study.
The producer of this study, DNV GL – a company that provides classification and technical assurance along with software and independent expert advisory services to the maritime, oil & gas and energy industries – is hoping their tools and techniques will lower the 15% failure rates. Since most products are only certified to UL and IEC standards – neither of which test long term performance – there is a clear market need. The Project Developer that best makes use of 15% lower failure rates will win the long game.
In the past 2 years DNV GL has executed 40 Qualification Programs across 30 manufacturers. The results of the study showed that overall, many module vendors performed well across all tests. For example, 8 manufacturers degraded less than 3% after 4 times the IEC duration in Thermal Cycling (the IEC pass/fail criteria for 200 cycles is 5% degradation). Two manufacturers performed in the top group on every test: Kyocera and Phono Solar. Roughly 55 – 60% of top group modules were manufactured in China. This is roughly equivalent to the ratio of Chinese module participation in the full PV Module Reliability Scorecard. This demonstrates that manufacturing location is not a good proxy for reliability. More importantly, that these manufacturers would spend money on these very hard tests means they see a opportunity for improvement, and we – the marketplace – will benefit.
As a Solar Power Project Developer during the day, I build financial models that take into account 25 years of electricity production. The industry standard 25-year Power Production Warranty that comes with a solar module was first introduced in 1997 – meaning the warranty itself is only 19 years old! If we’re going to get to 15,000GW of installed solar power – and we’re going to to trust solar electricity with powering our lives – then tools like this will be needed.