Proofing to avoid PV pandemonium

As published in PV-Magazine December 2022

The same product tested by different laboratories may return different characteristics. Mathieu Amendola explains the various reasons for this.

With each new module generation, we see more energy generated at lower costs. This trend looks set to accelerate with the development of high-efficiency cells such as TOPCon. Module performance characteristics, provided by manufacturers in PAN files, are improving and it seems that this is only the beginning. While everyone wants to hear that their project is going to perform brilliantly, we must take a balanced approach to manufacturers’ claims and cross check against sound data to make sure that “Peter-Pan” files do not turn your project into PV pandemonium.

One of the fundamental questions that comes to mind when considering a new solar asset is how much energy it will generate. Every kilowatt-hour counts, as solar stands on its own two feet and attains cost parity in many markets that have moved away from feed-in tariff or contract-for-difference models. This leads to long discussions between developers, lenders and technical advisers, with heated negotiations to ensure the best performance ratio.

Fifteen years ago, when the first large PV developments started, the art of performance simulation was still in its infancy, and it was common practice to apply generous safety margins when calculating the future performance ratio of an asset. As the years have passed, modelling tools such as PVsyst have improved, and confidence in the simulations have grown as the real-world performance of assets matched or exceeded the simulations. Real-life data were used to feed into and improve the models. A lot of work was dedicated to understanding how PV systems behave in the real world. Degradation parameters, soiling and temperature impacts were gradually better understood and modelled by the simulation software. There is no doubt that we now know more than in the past, so why should we still be concerned by the pre-construction energy assessment? Are our models and simulation software not good enough for what we need to achieve?

Predicting the future

It is true that the industry now benefits from years of operational data. Nevertheless, the technology continues to evolve and the products that will be deployed in the future present characteristics that are significantly better. Manufacturers have learned from the past and are adjusting their products to deliver more energy under local climate conditions. Two parameters are particularly intriguing: the low-light response and the angle of incidence modifier. By low-light, we mean the capacity of a module to keep a high efficiency under low irradiance (as low as 100 W to 200 W per square meter). Traditionally, high-efficiency cells under STC (1,000W/m²) will have relatively poor performance at low irradiance.

Most PERC and TOPCon cells show a very high performance at low irradiance. This, in theory, will allow the modules to produce more under cloudy skies for example. The incidence angle modifier reflects the fact that when the sun is low in the sky, a more significant fraction of the sun’s rays will be reflected by the front glass of the modules than when light hits the modules from a higher angle. Anti-reflective coating, texturization of the glass or of the cell’s surface reduce those losses.

Test labs

Proofing to avoid PV pandemonium
Researchers put modules through their paces in labs to create PAN files for use in energy yield simulations. The accuracy of these files is quickly improving, but discrepancies between those produced by different laboratories show that there is still work to be done.

It is common practice for module manufacturers to provide independent PAN file reports. PAN files are created by laboratories from samples provided by the manufacturers and follow the IEC 61853-2 standard. All laboratories will abide by this IEC standard, so why don’t we all live happily ever after? Well, the devil is in the detail. And the details can be cruel. The standard requires only three samples to be tested and not all samples need go through all tests. The low-light or low angle performance may be measured on only one module. Even a relatively small project, let’s say a 10 MWp ground-mounted project will see the installation of more than 15,000 modules – and they may not all behave exactly as the one tested sample did.

The same product tested by different laboratories may return different characteristics. The reason for this is the accuracy of the test equipment that will vary from one test site to another, even if all the materials are of prime quality. This was well illustrated by a publication of the 35th EUPVSEC conference, where eight leading European laboratories showed the comparison of the incidence angle modifier measurement they realized on the same sample cells. This publication showed significant discrepancies from one laboratory to another, in particular at high incident angles, with up to 20% difference from the most optimistic to the most pessimistic measures.

Too good to be true

The growth of our industry has been unstoppable so far and there is no reason to fear it will come to an end. On the contrary, every year that passes, no matter what happens, always ends with more new capacity than the previous one. It took more than 40 years to reach 1 TW of installed capacity. The second terawatt will almost certainly be installed by the middle of this decade. But as PV is becoming more competitive, we need to become even more professional to ensure a steady and healthy growth. We’ve learned from the past that mistakes can happen, we faced and handled new failure modes, and we will certainly discover new issues as time passes and new technologies emerge. We need to maintain some level of contingencies for the unexpected. And there is still a long way to go to keep improving the quality of our work as a profession. Current market practices must continue to evolve to maintain confidence in solar. In the end, what the world needs is a cheap and safe way to produce clean energy. We are already doing it, so let’s go further by remaining realistic and professional.

The same product tested by different laboratories may return different characteristics. The reason for this is the accuracy of the test equipment, which will vary from one test site to another.