Imagine the following situation: you have a large dataset you used to develop your doctoral Thesis. You publish a paper describing the phenomena in the long term. You answer questions and open others. So you go afterwards to find the answers based on previous knowledge – yours or from the state-of-art – and realise that what you are seeing is something different to what you were expecting. The first thing you think is that your suppositions or even the dataset have something wrong. At least until you start to dive into your notes, data, and theory, then you go back to the notes, theory, notes, data and so on.
The case mentioned here was the starting point of the paper recently published in Nature Communications Engineering entitled “Energy losses in photovoltaic generators due to wind patterns”. In a previous publication, my colleagues from then and I found a medium-term variation, year-by-year, related to the mismatch losses monthly evaluated. At that time, we did have not a satisfactory explanation for that. Suspecting that these periodic variations could be related to the wind patterns, I started to analyse the local wind measurements and how the temperatures would behave depending on wind velocity variations. For me, the clearest thing to do was to quantify the real impact of the wind on the energy losses in a large PV generator, considering that this impact was well-known in stand-alone PV/PVT. Now, observing for the first time this impact in a larger system, the thing here was to determine numerically how the wind speed increase would benefit energy production, based on the vast available literature about the benefits of PV cooling. Surprisingly, results showed an energy loss increase with the wind speed increase, assessed through the experimental mismatch losses analysis between PV modules. The first thing you can think is that something with your data went wrong. So, it is time to go back to the theory.
The temperature differences induced by the wind
First of all, we can imagine the PV generator as a flat plate exposed to a fluid – in this case, the wind. Imagine now the wind flowing parallel to the PV generator, changing its regimen from laminar to turbulent. Just at the beginning of the flow, in a laminar regimen, the heat transfer is high, decreasing fast until the transition zone, where the turbulence begins to develop. While the boundary layer thickens with the turbulent development, the heat transfer from the panels to the air decreases. As a consequence, the temperatures tend to be as higher as more developed the turbulence is. This gives place to temperature differences inside the PV generator, as greater as the wind speed is. Along the same lines, lower or even null wind speeds lead to lower temperature differences. Here I would like to point out two observations: 1. The temperature differences increase proportionally to the in-plane irradiance. The PV panels are more sensitive to wind incidence and consequently temperature differences for high irradiance values. 2. The temperature distribution remains the same even if the PV generator is not generating electricity, reinforcing the idea that the thermal behaviour inside PV generators must be dominated by fluid mechanics properties rather than any consequences of the Joule effect, for example. Additionally, this suggests that the losses induced by wind must occur in any kind of PV generator, independently of the technology in which it is made. The reader can argue that this is an ideal situation, as the PV generator is exposed to the natural wind, inherently turbulent. However, the observations presented here show that the temperature distributions behave similarly as expected in an ideal case (expected according to the theory), which can be well explained by fluid mechanics properties with great accuracy. It is worth mentioning that these considerations could potentially simplify analysis on a large scale, for example of a PV plant exposed to regional wind patterns. In addition, this same behaviour is seen even with the wind reaching diagonally the PV generator, both in the front and rear sides.
What about the benefits of PV cooling?
As we know how the temperatures behave with the wind incidence, the next step is to see how the mismatch losses behave in the same situations. Contrary to the expectations of energy increase with the wind speed increase, the mismatch losses increase with this increase. Along the same lines, the lower mismatch loss values are seen when the wind speed is low or even null. As many times told by many scientific publications, the wind can effectively bring some benefit concerning the PV panels cooling. This is somewhat correct, until a certain point. According to the results, the temperatures can greatly differ even in side-by-side modules, mainly if they are at the beginning of the air flux, due to the fast heat transfer decrease before the regimen transition. So, how many modules must exist to lose energy? Before you can call it a loss? A few modules, maybe just two.
A similar mismatch loss pattern occurs yearly. The higher values occur in the warmest months for this location, when a thermal low in the Iberian Peninsula inducts wind coming from the South and South-West quadrants. For this specific generator, south-oriented, the wind from these directions is responsible for the higher mismatch loss values.
For larger PV generators, as we can see, the story is a bit different.
Does this mean that the energy yield estimations were wrong?
No, absolutely. The point here is that the hitherto depreciated local wind patterns must be taken into account for a proper energy estimation during the lifespan of a PV power plant, despite the low value of the observed losses. With the PV module manufacture progress, the differences between them tend to be even lower. So each detail counts to reduce the uncertainties of the energy yield estimations. Besides, as many studies consider the benefits of harnessing the wind even in big installations, these results could suggest another path to take.
We still don’t know how this could impact the energy production from large PV power plants. Maybe the wind, as it induces thermal variations inside the PV generator, could increase the ageing of the modules. Maybe the wind patterns for a location must be taken into account during the lifespan of a PV plant, as they tend to change during the decades with the global climatic change. There are things that we still don’t know, and others that it could take longer to know. The answer is blowin’ the wind.