Striking the right balance between the front and rear side of a solar module, to maximize power yield, is no small feat. The sheer mass of parameters that go into the calculation can be mesmerizing. Starting with ground albedo – which changes over the course of a year – and extending to ambient and module temperature and diffuse light conditions, the list goes on.
Klaus Hofmeister, product manager for Europe at Trina Solar; and Kyumin Lee, engineering manager at CFV Solar Laboratories, shared their insights during a pv magazine webinar. With much left to talk about, our experts took the time to answer another batch of questions from the audience.
Are the new algorithms supposed to be installed on trackers available in the market or only Trina trackers?
The optimized tracking algorithm will only be available with Nclave trackers provided by Trina Solar.
Reflected radiation depends on the type of ground and its geometry, which is irregular. How is it possible to keep in consideration all these variables in an algorithm for each site? Wouldn’t it be easier to use sun direction sensors to measure the radiation on the rear side and control the tracking?
The algorithms use sensors on the front as well as on the rear side to control the tracking but even then it is difficult to really anticipate all operating states of a single tracker. In order to optimize output even further, the tracker-to-tracker dependency is also considered, as well as different mounting heights of the trackers. That certainly adds complexity to the system and requires a variety of input parameters.
One-year albedo measurement could be a long time for a PV plant project. Is there any approach to estimate one-year albedo based on a short time measurement?
Determining the albedo does not require a one-year cycle. Anyhow, it is very difficult to determine the albedo of a surface because it might change with the seasons; for example, grass will grow higher or maybe in the winter season the grass is covered with snow. For simulation, I would expect that an annual average assumption for the albedo is the best approximation.
There are efforts to correlate short-term, on-site measurements with satellite albedo data, with the hope being that you may ‘calibrate’ the satellite albedo data with short-term, on-site measurements and correctly capture the seasonal effects from the satellite data. This approach seems to work okay on uniform terrains such as desert, not when there is terrain variation within the satellite data ‘pixel’. In such cases, the approach would be to start with a ballpark figure from the satellite data and then adjust it as the project progresses and the on-site measurement data becomes available.
Could you explain how is it possible to calculate maximum and minimum voltages for bifacial modules? In other words, temperature vs. voltage?
The calculation of the voltage isn’t much different from a monofacial module. Bifaciality does not mean that the voltage will increase; it is mainly current, which follows the front and rear side irradiance. A bifacial cell still only has one p/n junction.
The PV modeling software includes the thermal model calculation, the incident irradiance on both sides of the PV module. So the higher temperature due to the rear side irradiance is already accounted for in the modeling calculations.
Is there a guideline on how to treat temperature losses and thermal dissipation factors in PVsyst for bifacial modules?
Currently there is no difference in the settings between monofacial and bifacial.
For now, it is safe to assume an identical thermal model for bifacial and monofacial PV modules, although there are some reports of the glass-glass modules actually operating at lower temperatures than monofacial glass-polymer modules for the same incident irradiance. Over time I expect module manufacturers to provide guidelines for the thermal model coefficients (Uc and Uv in PVsyst), based on actual testing.
Could you please explain how active and reactive power measurement help improve tracking?
In the current product development phase this is not yet implemented. At this point we won’t be able to provide further details on this topic.