Watching Apollo in action in its dedicated darkroom in the European Solar Test Installation (ESTI) in Ispra, Italy, puts a whole new sizzle on PV calibration: shutters slowly swing open revealing an array of 11 zenon lamps while several feet ahead a curtain draws aside. The air conditioning system provides the soundtrack, which is on at full blast to ensure temperature stability. The Apollo AAA large-area steady-state solar simulator (from Taiwan’s All Real Technology Company, the only one currently in use in Europe) is one of several new simulators at ESTI’s facilities, inaugurated this June following a €3 million refurbishment.
ESTI is part of the Institute of Energy and Transport, located within one of the European Commission’s Joint Research Centers. It is one of Europe’s primary reference laboratories with ISO 17025 accreditation for calibration and verification of PV devices (usually modules but occasionally verification of new prototypes). “At ESTI we focus on traceability and the reliability of calibration through the PV value chain,” says Ewan Dunlop, Technical Manager at ESTI, during a facility tour. He describes how this extends from government ministries that develop renewable energy polices, through testing labs and manufacturers to the end user.
Most ESTI services fall into two broad areas: electrical measurement and calibration of PV products and lifetime testing and module inspection for “type approval” testing to International Electrotechnical Commission (IEC) standards. Type approval testing confirms the quality and electrical safety of PV modules and is required in markets that offer incentives like feed-in-tariffs. This is done either in climatic chambers for accelerated aging and stress tests or in ESTI’s outdoor test field to check the impacts of environmental stressors like rain, frost, wind, snow, sand, temperature variance, etc.
PV calibration tackles the issue of measuring sunlight consistently. How can a kilowatt of sunlight be measured? If a ruler were used, would everyone agree on what constitutes an “inch”? Once irradiation is converted to amps, the power of PV modules is measurable. Calibration (for photovoltaics) relates readings (of solar irradiation) obtained by a measurement instrument (radiometer) to one of seven measurement units (amperes) for one of the seven base quantities (electric current) in the International System of Units. In the absence of absolute standards for irradiation measurement, national physical labs and metrology authorities collaborate on comparative standards. “From the client perspective, calibration offers them confidence in their data specs,” says Dr. Werner Herrmann, Senior Expert and Team Manager, Solar Innovation/Large Scale Projects at TÜV, a German testing and certification organization.
Calibration of the “primary reference cell” is the first and most challenging step in the calibration process. The objective is to create a measurement of solar irradiance to set the irradiance level of solar simulators. ESTI is one of the world’s four recognized primary reference cell calibration labs; accredited for three methods (global sunlight; direct sunlight; and solar simulator). By using two types of radiometer (a cavity radiometer and an absolute spectral radiometer), ESTI can evaluate spectral irradiance and do primary reference cell calibration at a level similar to that of a national physics lab.
Notwithstanding their diminutive size, (2cm x 2cm) ESTI’s set of five primary reference cells is invaluable and irreplaceable — some have up to 20 years of calibration history. Therefore, ESTI develops “working standards” from these for daily use. These are used with solar simulators to develop “secondary references” for industry or other test labs. Industry uses “secondary references” to test the current-voltage characteristics of PV modules under a set of Standard Test Conditions. A “traceability chain” (a documented unbroken chain of calibrations following the measurement process from start to finish), provides an estimation of the uncertainty associated with measurements at each stage.
Herrmann describes how secondary references are used in industry: “Since we are a testing and certification body we do lab testing of modules and conduct periodical factory inspections to ensure that production corresponds to lab testing. These inspections also include assessment of measurement practices at production sites,” explains Herrmann. “The experience of the PV module manufacturer is an important factor. Since PV module designs are frequently changed [as manufacturers continuously seek to optimize cost and efficiency] one secondary reference module may be used for a certain range of products. It’s not practical to have a reference module for each module type. However if optical transmission properties or cell type is changed, the manufacturer must decide whether to change the secondary reference,” he adds.
Dunlop details how different types of simulators have different functions. ESTI normally uses the Apollo to develop secondary references and occasionally to verify a sample of PV modules for a large purchase contract. Since it can provide full sunlight conditions over a 2m x 2m test area for up to 8 hours, Dunlop explains that the Apollo is especially useful for new high-efficiency PV modules that are challenging to test for several reasons, e.g., an ultra-high doping content. The high capacitance of high-efficiency modules renders them difficult to characterize in a single flash. Pulse simulators may give false readings since some modules require time for preconditioning.
While Apollo takes 30 minutes to power up, ESTI’s Pasan IIIB large-area pulsed simulator (3m x 3m) is energy-efficient and provides a constant module temperature. A series of diaphragms improve spatial uniformity of the optical chamber so as to avoid creating hotspots while measuring electrical performance. “We use it to calibrate factory references for manufacturing of standard mono/polycrystalline silicon modules,” says Dunlop.
Other new acquisitions include a high-intensity simulator from Canada’s Sciencetech, which can simulate the power of 200-2000 suns over 4cm x 4cm. “We support the IEC’s Technical Committee 82, Working Group 7 on the development of CPV related standards. You must know the spectral content of the light that you’ve measured under to know what you’ve measured,” says Dunlop. “The Spi-Sun on the other hand is a solar simulator typically used in industry. Because of its design — like a desk — it’s useful in line production since a module can roll onto it. We use it to check the transfer of calibration made on the Apollo or Pasan to the industrial measurement environment.”
While metrology rarely makes headlines, the calibration of reference devices and subsequent verification of PV technologies and type approval of products are fundamental to the economics and bankability of the PV industry. Herrmann estimates that assuming a nominal module price of approx. 0.60 €/W, every ±1 percent measurement uncertainty on peak power corresponds to a value of > ± €216 million (assuming annual installation of around 36GW for 2013).
“The purpose of estimating measurement uncertainty through the calibration traceability chain is to minimize the potential for systematic error at production sites, which can be very costly,” he continued, cautioning that all calibration refers to one test lab. “In reality the range of measurement uncertainty is more likely to be ±2 percent for crystalline silicon PV modules due to variability among labs as determined by round robin performance testing. Test labs are collaborating to increase comparability.”
Given low cost margins on PV modules, profitability is sensitive to the reliability of electrical performance measurements and the associated uncertainty. Whether for customers, manufacturers, insurance companies or investors, accurate measurements of performance and efficiency, and the interactions between stress-testing and performance, will be critical industry sustainability. In the end, says Dunlop, “Here’s what it comes down to: is the customer of the PV module getting what they paid for?”
