Array Technologies’ DuraTrack system is much different than Soltec’s (and most other trackers really) with its centralized drive, longer rows and its one-module-in-portrait configuration (1P) versus the two-in-portrait approach highlighted by Soltec (2P).
Array’s testing approach combined a mix of predictive modeling software and real-world testing so that each would validate or inform the other. For starters, they used the 2D modeling of PVsyst, an industry standard approach, but one that is limited in grasping the gains and losses of backside irradiance at this early stage of our understanding of bifacial performance.
To better inform PVsyst in input fields like “Structure shading factor” and “Mismatch loss factor,” Array partnered with PV Lighthouse of Australia, which specializes in a state-of-the-art 3-dimensional raytracing. This approach calculates all relevant physical effects, “from macroscopic structural shading and reflection down to microscopic cell texture and quantum efficiency of solar cells.”
Finally, to validate the PV Lighthouse modeling studies, Array carried out field testing at CFV Solar Test Laboratory, an ISO 17025-accredited test laboratory in Albuquerque, N.M. Important note: only the inner rows were used for data collection, as Array felt this more accurately reflected the effect of periodic ground-shading patterns in utility-scale installations. Shading and mismatch losses tend to vary with the time of the day and the weather condition, but for the study PV Lighthouse determined best-fit values to represent yearly averages, and then those values were input for both 1P and 2P trackers in their typical configurations in a sunny climate. The albedo was ~0.3 over the test period.
“Small-scale field tests are not adequate to yield accurate results,” Array notes in its report on their methodology. “This is because bifacial gain numbers change significantly from a single-row system to periodic long rows that are more representative of reality. Undersized test setups in many previous studies have unfortunately resulted in inflated bifacial gain numbers that are not realistic for utility-scale installations.”
You can (and should) check out the company’s in-depth findings in this white paper, but here are the main takeaways.
Backside irradiance is erratic
There are three sources of backside irradiance to try and capture. The main source, and the most often considered, is the ground reflection (albedo). But Array’s study also considers the diffuse irradiance arriving directly from the sky, and the reflection off the front side of the PV modules in other rows.
“One interesting result was that the structural shading factor was not too different for the 1P and 2P tracker configurations despite the ratio of the torque tube width to the PV row width being smaller for the 2P tracker. This was due to the ground-reflected rays arriving at a more oblique angle to the back of the PV rows for the 2P tracker than the 1P tracker on average, due to the lower aspect ratio (row height over row width) of the 2P tracker.”
This was an important finding because plugging the values into PVsyst calculations with the 2D view factor model showed to be fairly accurate.
Hot spots not an issue
If a cell doesn’t produce enough current to match the operating current of the module array, a hot spot can occur, and this is a big concern of bifacial skeptics — understanding the impact of cell-to-cell irradiance mismatch turning into a hot spot.
Array put what it called a worst case scenario in place to try and force mismatch:
“Artificial white ground cover (albedo > 0.6) was installed below and around a PV module array installed on the south end of the tracker installation at CFV. The tracker was intentionally parked at its maximum angle to the east, to simulate the automatic wind positioning of Array Technologies’ trackers. As the sun moves to the west, the contribution of the backside irradiance to the total irradiance increases, maximizing the effect of structural shading on the cell-to-cell irradiance mismatch.”
Neither the IR thermal images of string voltage data showed any signs of BPD activations.
“We believe this test is a good representation of what can happen under snow-fall conditions, for a split-cell bifacial module array mounted on Array Technologies’ 1MIP tracker. As we did not observe any BPD activation on the setup as tested, we think the risk of bifacial PV arrays suffering hot spots due to structural shading is minimal for our trackers.”
Row width vs. height
Somewhat surprisingly, these tests showed that 1P trackers have more bifacial solar resource coverage than 2P, with the difference being more pronounced in snow cases. To make sense of how they achieved these results, Array used back-to-front irradiance ratio (BFIR; ratio of backside irradiance to frontside irradiance).
“Upon analyzing these results in depth, it was discovered it is not the row height that drives the BFIR; it’s actually the aspect ratio (row height divided by row width) of the system. 1P trackers achieve higher BFIR than 2P trackers because its geometry allows them to capture more of the ground-reflected light. The issue with 2P trackers is that their height normalized to the row width (=aspect ratio) is actually lower than 1P trackers, resulting in wider windows through which the ground reflected light is lost to the sky.
“With this new insight, we replotted the results of the PVsyst study, this time using the aspect ratio as the x axis (FIG 12). Now the 1P and 2P results merge seamlessly, which shows that the aspect ratio, and not the absolute row height, is the correct metric for the bifacial tracker geometry.”
— Solar Builder magazine