PHOTON International The Photovoltaic Magazine

Ispra's Heinz Ossenbrink about energy rating of PV modules

The IEC is working on the draft for a new standard called IEC 61853 »Performance testing and energy rating of terrestrial photovoltaic modules,« which will define uniform guidelines for the calculation of both nominal power and attainable yields. After the »Module Yield Workshop« in Zurich, PHOTON International's Jochen Siemer spoke with Heinz Ossenbrink, head of the test lab at JRC Ispra about the new standard.

	© Heinz Ossenbrink
Heinz Ossenbrink has worked since 1982 at the Joint Research Center (JRC) of the European Commission in the northern Italian city of Ispra. For the past five years, he has led JRC's Renewable Energies Unit. About 40 of JRC's almost 1,400 employees are developing measurement and test procedures for harmonised standards of PV components.

PHOTON: At the »Module Yield Workshop« in Zurich, one participant suggested after a lengthy discussion that perhaps people should accept the fact that working with solar energy has a great deal to do with nature, and so not everything can be predicted from the outset...
Heinz Ossenbrink: I fully agree...
PHOTON: But making predictions is exactly what module manufacturers and system planners would like to do. That's why they are working to develop a standard that will make such calculations uniform.
Ossenbrink: Currently, competition between solar modules is based primarily on the dubious watt peak (*) - and power is generally the key selling-factor. But users would really like to know what the result will be, for example not the »watt peak,« but the »watt-hour.« So the next step is to allow for comparison of modules according to kilowatt-hours. Obviously, one cannot standardize such a comparison for use in every location on the planet, each with its specific climatic conditions. Though the Standard Test Conditions (STC) for power measurement constitute a relatively artificial definition, they can still be used worldwide. That's our position regarding the Energy Rating, as it is defined in the draft IEC 61853: We need something that can be used to compare products, something that is reasonably measurable and determinable. It cannot be a number that will tell users how many kilowatt-hours to expect from the modules in their installations, since that depends on weather conditions, which as we know are unpredictable. But it can be a number that tells users: »This module will produce a certain yield under certain conditions.« Then users can compare products with one another, and that is our only goal.
PHOTON: Currently, the IEC 61215 standard and the IEC 61646 standard for amorphous modules are by far the most respected. In your opinion, is that a legitimate development, or are there other standards that deserve more attention?
Ossenbrink: First of all, one needs to examine what exactly a standard defines. The IEC 61215 standard is intended, insofar as it is possible, to ensure that a certain finished module can withstand extended use - though, naturally, the question immediately arises as to what »extended use« entails. So it says something about the lifespan of the module. Often, this causes a bit of confusion, as the test reports also include the power data of the modules. Those data are reported, however, to document any changes that arise during the various module tests, i.e. »During test x we observed a change in power of 6 percent.« In a formal sense, then, the IEC 61215 standard is not a certification of quality, but a certification that a module type's construction can last 20 years and deliver certain power.
PHOTON: What other types of questions does a certification of quality need to answer?
Ossenbrink: For instance: Once manufacturers have received certification, will they continue to use the same materials? Will they build their modules using the same procedure? To define something like this is the task of quality-control systems that will eventually become the standard for solar modules. They must be able to clarify the most difficult points that we are currently unable to cover.
First and foremost, that refers to the selection of modules to be tested. Although the IEC standard calls for the modules produced through normal methods, we cannot verify this. Then there is the question of how long the certification should be valid. Over the last ten years, we have tried to promote ISO 9000 and other similar standards for industrial quality systems, whereby one would say: »Okay, you've developed a good product, it has certification - but now you have to prove you can maintain this level of quality.«
PHOTON: Safety questions are very important, as are questions of quality, naturally, but developers and customers are primarily interested in yields. That's the goal of the IEC 61853 draft: a norm for yield calculation. Is this the big thing the industry has long been waiting for?
Ossenbrink: I believe so. The draft presently defines five standard days, including temperature profiles. Thus, one can have five different energy yields, and in my opinion that's still too many. But as things are envisioned, one could then take these five numbers and arrange them on a single label. A dealer could ask customers where they intend to install their system. If they plan to use it in the tropics, they should use the number for »warm and sunny.« If they are headed to Norway and Sweden, they should take »cold and sunny.« Then customers can ask other manufacturers whether they can get a module better suited to »cold and sunny« at the same price. That is certainly a major advance. We will have to wait and see; standards like these are revised every few years to ensure that they prove successful.
PHOTON: How does one measure that success?
Ossenbrink: I am personally of the opinion that one can manage such an operation fairly easily. And according to the IEC 61853, for the daily profile I only need to take 12 measurements for irradiation and 12 for temperature and then combine that information with the module data. One doesn't necessarily need to take new measurements, since the necessary data, such as temperature coefficients for voltage and current, are already included in the existing standards. So we're a big step closer. In the current draft of the standard, there is a suggestion to take into account also for sunlight's temporal spectral changes that will make the calculations more comprehensive, but would also allow a better comparison of different PV technologies.
PHOTON: So no additional measurements are necessary for the calculations?
Ossenbrink: We hope to make it possible for developers to do this with little money and equipment, as it is a bit of paperwork. Also, the test laboratories will not require further investment. And the industry will pay almost nothing extra, as they can do the calculations on their own.
PHOTON: Though always assuming that in test laboratories measurements are carried out correctly and that the industry adheres to the quality standard...
Ossenbrink: Of course. I will assume that both test laboratories and the industry are adequately following literally of the existing standards.
PHOTON: Let's return to the most interesting example, rated power. An enormous number of modules with a watt rating of +/- 10 percent are being sold, although they only adhere, if even that, to the lower values of this range -- and only with great effort. Thus, a 100 W module with just a 91 W power rating still fulfills the IEC 61215 standard. Shouldn't there be a regulation that requires a declaration of the watt rating?
Ossenbrink: We are currently working on such an amendment; it has the quite amusing title of »Blank Detail Specification.«Essentially, there are two different philosophies at work here. One calls for a pure test standard such as the IEC 61215, which actually only describes procedures according to which the finished modules are to be tested. The other envisions quality control for the entire production process. The well known problem of declaring +/- 10 percent has given us a confirmation of just how exact the industry's measurement can be; in fact often modules have a 10 percent negative tolerance +/- 1 percent. The problem of the industry is that solar cells alone have a production spread of around +/- 20 percent. Certainly they are sorted and placed in an appropriate group, and then it is up to the manufacturers whether to reveal that to customers. Many manufacturers have various module subtypes, so they can ask customers: »Would you prefer our 105 W module, or is a 90 W version sufficient?« Others simply use a medium-sized module and sell it at +/-10 to 20 percent. The Blank Detail Specification will help to limit this practice. I sent off comments on the draft four weeks ago.
PHOTON: How exact will the power ratings be?
Ossenbrink: The manufacturer will be required to guarantee +/- 5 percent. The problem with production spread that arise during manufacturing -- which is not a question of quality, but an unavoidable side effect of using silicon -- can be declared by using different module types, i.e. types A, B, C. In the semiconductor industry, such declarations exist in the form of various product classes, for which one receives a guarantee upon order. We could thereby solve two problems with solar modules at the same time: The user wants the most accurate specifications possible, and manufacturers, which usually issue 20-year power guarantees for their products, still want to ensure that they can sell all of the power variations that arise during manufacturing. We want -- and many dealers would be quite satisfied with this -- to allow a module to retain its name (XYZ), but be offered with a subtype, i.e. XYZ 100-A, B, or C, which would stand for a corresponding power rating.
PHOTON: Thank you for the interview.

(*) Watt peak: The greatest possible power of components or systems. PHOTON articles generally refer only to the unit »watt.«