Bifacial PV technology: technical considerations
Bifacial PV converts sunlight to DC electricity on both the front and back of the PV modules, and it can be used about equally well with either a fixed-tilt or a tracking structure. Figure 1 shows a general view of a bifacial PV module mounted on a single-axis tracker and the routes of the sun rays around a bifacial structure is also showed.
Bifacial PV technology has been deployed in limited quantities since the 1980s but has only begun to capture significant industry attention since approximately 2016.
Figure 1: Views of bifacial PV (images by courtesy of Soltec)
A bifacial PV plant generates more energy than any monofacial or conventional approach with a variable percentage also called bifacial gain. The Capital Cost and O&M costs associated to the bifacial are also higher than the conventional technology. Therefore, the key element is to know about the bifacial gain. Is there a rule of thumb to be used?. It would be very risky applying a gain with a supported design and calculations. Real bifacial gains reported vary with many factors on the PV plant design and the location. Therefore, applying a single bifacial gain fitting all projects would be not reasonable.
The yield from a bifacial PV system is dependent on the same parameters as for standard PV systems, such as ground cover ratio (GCR), orientation (tilt and azimuth or tracking angles), and meteorological conditions such as irradiation quantity and quality (spectrum), diffuse fraction, temperature, wind speed, albedo, and dust and snow deposition. However, certain of the same terms take on added importance for bifacial, such as the albedo and the GCR, and a couple of new factors come into play that are largely unimportant for standard PV, such as the front aperture ratio and the degree of structural sunlight blockage on the back side. Let’s discuss a bit more in detail these factors together with the known performance degradation of conventional PV modules compared to bifacial PV modules.
Albedo and bifacial PV
Along with pure climate-related terms such as the quantity of solar radiation, temperature, wind speed, and precipitation (in the form of water and snow), there are other site-specific properties that will have a strong influence on bifacial performance. The most notable of the site-specific terms of interest for bifacial modelling is the ground reflectivity, or albedo. Albedo is meaningful even for monofacial PV when the tilt angle affords enough of a view factor for reflected light to be a significant supplementary contributor to the irradiation absorbed within the front-side plane of the array. This is especially beneficial in snowy, high latitude climates, where steep tilt angles are common. For bifacial applications, making use of ground-reflected light is a necessity rather than a supplement.
While most natural surfaces tend to be better absorbers of light than reflectors, with albedos that typically hover in the 15-30% range, there are natural materials that exhibit moderate to high reflectivity (e.g., water, snow, white sand). There are also prepared surface options, such as crushed white stone, that can be used to enhance reflectivity. The albedo of a site can vary significantly throughout the year, most notably due to intermittent brightly reflective snow cover in the winter, but also due to soil moisture fluctuations and vegetation. Within a multi-MW site, notable variations in albedo can occur from differing rock and soil composition and plant growth, so it is important to quantify site albedo as part of the bankable design process.
Albedo is not well-mapped, and as noted above, can vary significantly across a site depending on topography and climate.
As a coarse estimate of the sensitivity of energy gain to albedo, any fractional change in albedo ranging from 0 to 1 will introduce a corresponding change of about one-fourth that much in terms of annual energy gain or, equivalently, in terms of net capacity factor (NCF). For example, the fractional annual energy or relative NCF gain is ≈ (0.25) x (albedo). The difference in NCF between a surface with a 30% albedo, such as brighter sand, versus a dull dirt or dry grass ground cover with a 20% albedo, translates to approximately a 2.5% change in NCF.
The other site-specific properties that should be examined, as with any PV project, would include terrain slope/roughness and perimeter and horizon shadowing obstructions. Lastly, bifacial boost is expected to improve with latitude (assuming typical installation tilt angles) even though overall solar resource will decrease rapidly with increases in latitude. Spectrum, while normally a relevant term for any form of PV conversion, is not apt to be a differentiator for bifacial modelling, as spectral composition for bifacial does not seem to be viewed as significant in the current literature. While DNV agrees that there is a spectral difference between reflected light and direct light, the impact is small and difficult to estimate without further research in this area.
Bifacial PV plant design
As PV costs have dropped so dramatically over the past decade, higher GCRs have increasingly become the design norm. The tolerance for shading has increased significantly, especially in space-constrained settings, where more revenue can be gained by increasing the density of PV capacity per hectare, even at the expense of reducing the unit specific yield in kWh/kWp due to higher shading losses. For bifacial PV, high GCRs are not desirable, though, as a significant portion of the back-side radiation gain is inhibited by closely spaced racks that block northern sky radiation (in the northern hemisphere). So, for bifacial PV, optimizing value means introducing at least one design trend that is at cross-purposes with the design trend that leads to optimal standard PV systems.
Clipping loss and DC/AC ratio is another design aspect of bifacial that runs counter to the prevailing design trend in standard PV. Bifacial gain is most prominent at times when a system is at or near its maximum output, so an AC capacity limit may negate much of the value that a bifacial array might otherwise be able to deliver. The countermeasure is, of course, to reduce the DC/AC ratio and the propensity for clipping, with a small trade-off being an increase in inverter cost and in related dc BOS costs for equipment such as DC combiners and wiring.
The aperture and structural losses associated with bifacial will affect the performance throughout the system’s lifetime. Aperture ratios of at least 0.5 (module height above grade relative to plane of array row width) are recommended due to the sharp decrease in bifacial gain that can be seen for aperture ratios less than 0.5, with values of at least 0.75 preferred in order to begin to smooth out and increase the raw value of back side radiation. Non-uniformity of radiation on the back side, whether caused by proximity to the shaded ground or by direct blockage from structural elements, increases electrical mismatch loss within the cell string of the module and significantly cuts into the potential bifacial power boost.
Performance degradation
For long-run operation, an annual degradation of 0.5-1% has been presumed to cover about half of the potential long-term performance outcomes being seen for modern types of conventional PV. The magnitude of annual degradation is small and has historically been difficult to measure accurately. This is partly because of the almost undetectably slow rate and the limitations of data acquisition sensor accuracy. The large uncertainty is also attributable to many confounding factors that can either hasten or mask the slow annual decline, such as curtailments or any form of partial or full outages, including clipping.
For bifacial PV, the emerging discussion, at least qualitatively, is that bifacial technology, while lacking long field data histories, is not a fundamentally different technology leap in terms of observed degradation mechanisms and therefore should exhibit a degradation rate comparable to conventional PV. Proponents have argued that the internal cell architectures are very similar to well-established PV wafer technology, while sceptics have argued that introducing more internal cell contact points that are less insulated from outdoor extremes will translate into more defect mechanisms and a lower reliability and higher degradation rate.
To that point, the glass on glass construction used for some bifacial modules has brought up concerns about the inability to allow outgassing by-products from the encapsulant escape, as they are able to do with a standard polymer back sheet. However, some have substituted the common EVA encapsulant with polyolefin, which reportedly is less prone to outgassing as it cures and ages. DNV is not aware of deep studies concerning bifacial degradation. Therefore, assuming a long-term degradation for bifacial technology is another challenge faced by the engineers applying this technology.
Finally, DNV has made a considerable effort to improve the accuracy of the long-term production estimates from bifacial PV plants by developing SolarFarmer. SolarFarmer is a software tool already commercial able to simulate bifacial PV plants and it is also prepared to simulate modern single-axis trackers with new backtracking algorithms for complex terrain (individual tracker position every thirty seconds, for instance) and production improvement by setting the position at zero degrees when cloudy maximizing the irradiance on the plane of array. A screenshot of our software is showed in figure below.