A curious thing happened in the solar market in 2010. Relativelyexpensive modules from well-established manufacturers sold well, whileless expensive modules from startup firms did not sell as well. Why diddevelopers pay more? Modules from the established firms had demonstrated excellent, predictable energy production over many years. In contrast,startups had not built this data set for their modules. Establishedfirms have the financial strength to replace the modules if they failed. Startups had not yet established this strength. Developers paid more to reduce the risk to their investors — Banks.
Applied foresaw this "bankability" issue for our thin film silicon customers. We started an effort at the beginning of 2009 to collect the performance and reliability data demanded by developers and theirbanks.
Our first step was establishing test sites around the world. Avariety of regions were selected to show the module performance indifferent climates. The sites monitored are in Neustadt bei Coburg,Germany (cold/hazy climate), Singapore (hot/hazy), Phoenix, Arizona(hot/clear), and Santa Clara, California (moderate). For comparison, crystalline silicon modules were also installed in Singapore and Phoenix. The sitesincluded single junction amorphous silicon and tandem junction amorphous silicon/microcrystalline silicon modules.
The most remarkable result was the comparison between thin filmsilicon modules and crystalline silicon modules. In both of the hotlocations, Phoenix and Singapore, the thin film silicon panels produced>8 % more kWh per rated kWp than c-Si modules. Two notable trendsexplain the energy harvest advantage:
- Less negative temperature coefficient which gives thin film siliconmodules a performance advantage over c-Si modules at increasingly highirradiance and cell temperatures.
- At low irradiance, the thin film silicon panels out-perform c-Si,due to a combination of spectral and angle-of-incidence effects.
The performance data allowed us to establish predictability. Weproduced PVsyst models which predict the output of thin film siliconmodule arrays. The model predictions were compared to the actualperformance of the four sites. The actual performance exceeded thepredicted performance by ~5 %, which gives banks confidence that theycan calculate the financial output of an investment in an array of ourcustomer’s modules.
There is no fast way to establish 20-years of reliability data, butwe can run accelerated tests to show that the module design is sound.The reliability test lab in Xi`an, China has subjected thin film silicon modules to two and even three times the environmental stress than the standard tests require, and the moduleshave maintained performance within the limits set in theseinternationally recognized standards. This gives banks confidence thatthere will be no unexpected long-term degradation of the modules.
This data also allowed our customer to address the issue of financial strength. Our customers are using the data to attract insurance tocover their modules in the event of major failure. Banks know that there is someone standing behind the module manufacturer.
This data allows us to answer the big question: How does the cost of a kWh produced by our customer’s thin film silicon panels compare toother sources of electricity? How close are we to "grid parity?"Electricity production has a complicated pricing structure which depends on, among many things, location. The most expensive source ofconventional electricity is natural gas peak production plants, whichmeet demand during hot summer days. Riyadh, Kingdom of Saudi Arabia as a good location for thin film silicon due to the advantageous temperature coefficient, so we have compared thin film silicon modules to peakerplants in Riyadh.
To compare difference sources of electricity production, we havebuilt a Levelized Cost Of Energy (LCOE) model which calculates the costper kWh of a source of electricity over the entire life of that source.The model predicts that thin film silicon modules have an ~1 ¢/kWh lower LCOE than a comparable c-Si installation at an equivalent system costin Riyadh. Our present megawatt reference design, which we have madeavailable to customers, has a Balance-of-System (BOS) cost of 1.35 $/Wp(AC) for 9% thin film silicon modules. With thin film silicon modulemanufacturers continuing to drive to $1/Wp module cost, we should seesystems with total costs <$3/Wp (AC) in 2011. This cost correspondsto an LCOE of ~13 ¢/kWh, which is comparable to natural gaspeak-production power plants in Saudi Arabia. Our customer’s thin filmsilicon panels are at price parity for the peak production in thislocation.
We have established the bankability of our modules by demonstratingthe performance, predictability, and reliability of thin film siliconmodules. As a result, larger numbers of our customer’s installations are getting backing from banks. To reduce construction risk, we developedmounting structures, 1000V electrical systems, and shipping racks forthin film silicon modules. We demonstrated these designs at our campusin Santa Clara, California. For more information on the study which wepresented at EUPVSEC or our system designs, please feel free to contactme via this blog.