Search documents

Browse topics

Document details

Bifacial Performance Optimization Studies Using Bifacial Radiance and High Performance Computing
J.S. Stein, M. Prilliman, C. Stark, J. Nagyvary, S. Ayala Pelaez, C. Deline
PV Systems and Storage – Modelling, Design, Operation and Performance
Subtopic: Design and Installation of PV Systems
Event: 36th European Photovoltaic Solar Energy Conference and Exhibition
Session: 5CV.3.32
ISBN: 3-936338-60-4
0,00 EUR
Document(s): poster


Bifacial photovoltaic cells and modules are rapidly becoming recognized as a very likely and promising technological pathway to lower LCOE and higher energy yields. Recent market forecasts predict that the rapid increase in mono PERC cell production (e.g., Longi) along with new offerings in racking, trackers, and other BOS equipment designed specifically for bifacial systems will lead to rapid bifacial PV growth in the utility-scale market within the next two years [1]. Other cell pathways to bifacial (e.g., HIT and nPERT) are also showing promise due to recent innovations. Adoption of bifacial PV into the mainstream presents both a great innovation opportunity for expanding solar power, but also a major challenge for system integrators who lack sufficient modeling tools and design guidelines to fully take advantage of this new technology. We have learned that bifacial performance is sensitive to factors that have little to no influence on monofacial performance. For example, bifacial system performance is significantly affected by the height above ground, the dimensions and geometry of the module frame, mounting clips and racking, the configuration and orientation of modules in the array (e.g., 1 up portrait vs. 3 up landscape), and system configuration (number and length of rows, module spacing, etc.). Bifacial arrays are also affected by factors that affect monofacial arrays such as tilt and azimuth, ground albedo, and row spacing, but the effects are different for the bifacial arrays. For example, while albedo does affect ground reflected irradiance hitting a tilted monofacial array, its effects on bifacial arrays are much stronger [2]. Simulations models are being developed and validated to assess how system design parameters affect bifacial PV performance. There are two main classes of models that aim to predict the light reaching the back of bifacial arrays (1) view factor models and (2) ray-tracing models [3,4]. View factor models run quickly but usually require that the scene is simplified to only include effects in two dimensions (no edge effects) and also neglect the effects of racking, frames, and other nearfield objects [5,6]. Some 3D view factor modeling has also been applied to bifacial PV systems [7]. Ray tracing models (e.g., Radiance) are much more computationally intense but have the advantage that they can include details of the array and modules that may be important distinguishing factors for customers selecting bifacial modules for a specific project [8]. For example, framed vs. frameless modules, effects of different racking, gap between modules and torque tube, etc. This paper will report on work being done at Sandia National Laboratories and the Renewable Energy Laboratory to bring bifacial PV ray-tracing capability to a high performance computing (HPC) environment that allow high fidelity simulations to be run for time periods of interest (e.g., annual simulations rather than a few representative days) and conduct design optimization calculations, which require the ability to sweep through a range of input values to explore optimal module and system designs