Accuracy of ray tracing and view factor optical models for energy yield prediction of fixed tilt and tracked bifacial PV systems

Abstract: Bifacial photovoltaic (PV) modules capture light from both sides (front and backside) and convert it into electricity. This makes bifacial modules a promising technology in reducing the cost of electricity generation through the boost of the electricity generated by the bifacial PV systems over their lifetime in comparison to standard (monofacial) PV systems. The power output of bifacial PV systems depends on many additional factors in comparison to monofacial installations. Therefore, several simulation models for energy yield prediction of bifacial PV systems have been recently developed and are under continuous improvement. However, compared to the well-established models for monofacial PV systems, up to now, the accuracy of the models for bifacial PV systems has been validated based on field data to a much smaller extent. Accordingly, there is a higher uncertainty and risk regarding the reliability of the related yield predictions, leading in turn to lower confidence of end-users and investors regarding the profitability of bifacial PV technology and thus making it more difficult to finance and to implement large scale bifacial PV systems. Validation of these simulations models based on long-term field data acquired by test systems at different geographical locations and with different geometrical configurations, for both fixed and tracked
bifacial PV systems, are a prerequisite to overcome the aforementioned challenge. Furthermore, these validations will improve the reliability of the models, verify their assumptions, and determine their limits in practice.

This dissertation suggests a simulation model for energy yield prediction of fixed-tilt and horizontal
single-axis tracking (HSAT) bifacial PV systems. The simulation model is made of sub-models that were chosen rigorously based on recent literature review and related aspects for an accurate energy yield prediction of bifacial PV systems. The model deploys two main existing optical models (view factor and ray tracing) for front and rear irradiance modeling of bifacial PV systems. These optical models are combined with thermal and electrical models to allow for a prediction of the energy yield and the bifacial gain. The suggested model is validated using experimental data acquired on 6 PV systems with different mounting configurations and at different locations. Besides, various sizes of bifacial systems are investigated from stand-alone (200 WP ) to commercial-scale bifacial PV systems (0.4 MWP ). These bifacial PV systems are different in terms of installation configuration (such as mounting height, albedo, tilt angle, orientation) and the technology of bifacial modules.

The validation of the optical models is made possible through the fabrication of customized solar
panels for rear irradiance measurements in the field. The results from different experiments show a clear trend: independently of the size of the bifacial PV system, the measured rear irradiance tends to be overestimated by the ray-tracing model, while underestimated by the view factor model. On the other hand, the view factor model demonstrates better agreement to the measured rear irradiance for large size bifacial PV systems than ray tracing. Furthermore, based on the field data of one test system, the importance of taking into account the spectral reflectance of the ground underneath the bifacial PV system was demonstrated: in this way, the gap between measured and ray tracing modeled rear irradiance was reduced by around 8% relative.

Another topic that is investigated is the non-uniformity of the rear irradiance that is seen in fixed-tilt
and tracked bifacial PV systems. Thereby, it is found that the backside irradiance non-uniformity decreases with increasing tilt angles or increasing mounting height, and it is high during sunny days (high direct irradiance fraction), while on cloudy days the non-uniformity is reduced. The mismatch loss in power at the module level due to the rear irradiance non-uniformity was modeled by combining the optical, electrical, and thermal models. As a result, a quadratic relationship is found between the mismatch power loss and the non-uniformity level. The analyses show that - under typical conditions - the loss in power at the module level due to rear irradiance non-uniformity does not exceed 0.5% relative.

The measured power of the bifacial module located in the center of a PV array, surrounded by neighboring rows and modules, was compared to the power modeled by the view factor and by the ray-tracing model at different module tilt angles. The results have shown a good agreement for both models. However, the accuracy is found to depend on the tilt angle, with the vertical configuration being most difficult to model accurately. Overall, ray tracing models slightly better the irradiance - and consequently, the power - of the bifacial module for different tilt angles than the view factor model does.

Regarding the system size, it has been found that the 2D view factor model (assuming unlimited row length and requiring a lower computational effort) is suitable for modeling large scale bifacial fixed-tilt as well as single-axis tracking PV systems. Thereby, compared to measured values, the 2D view factor has shown a low deviation for long term (6 months) modeled energy yield and bifacial gain with an accuracy of around +/-1.5% (relative) and +/-0.5% (absolute) respectively. On the other hand, an array size of 3 rows with 7 modules per row has been determined as a lower limit for the application of 2D optical models. Accordingly, to accurately model small size bifacial PV arrays (not larger than 3 rows with 7 modules per row), a 3D approach is required for the optical model which - within the framework of the present Ph.D. dissertation - is implemented with ray tracing. A comparison between the model developed within the present work, the commercial software PVSyst and the model developed by ECN (now TNO) show a good agreement in terms of the modeled bifacial gain at different tilt angles to the measured data, which confirms the maturity of these three bifacial PV simulation models. In a case study, the simulation model developed within this dissertation was used to explore the potential of fixed tilt and tracked bifacial PV systems mounted in the Sahara Desert, the results have shown that combining bifacial modules with single-axis tracking leads to a significant increase in energy generation of about 36% in comparison to a fixed-tilt monofacial PV system. The results of this dissertation have made clear that the energy yield of bifacial and monofacial PV systems can be modeled with similar accuracy which gives more confidence and certainty to the modeled energy yield of bifacial PV systems

Angaben zum Objekt

• Standort: Deutsche Nationalbibliothek Frankfurt am Main
• Umfang: Online-Ressource
• Sprache: Englisch
• Anmerkungen: Universität Freiburg, Dissertation, 2020

Klassifikation und Themen

• Schlagwort: Bifacial-Solarzelle ; Simulation ; Fotovoltaik ; Energie ; Ray tracing

Beteiligte, Orts- und Zeitangaben

• Ereignis: Veröffentlichung
• (wo): Freiburg
• (wer): Universität
• (wann): 2020
• Urheber: Berrian, Djaber
• Beteiligte Personen und Organisationen: Glunz, Stefan ; Albert-Ludwigs-Universität Freiburg. Fakultät für Angewandte Wissenschaften

Weitere Informationen

• DOI: 10.6094/UNIFR/169842
• URN: urn:nbn:de:bsz:25-freidok-1698423
• Rechteinformation: Der Zugriff auf das Objekt ist unbeschränkt möglich.
• Letzte Aktualisierung: 25.03.2025, 13:48 MEZ