Modeling Partial Shading at the Cell Level on Photovoltaic Modules
No Thumbnail Available
Author
Calin, Jean-Paul
Faes, Antonin
Mujovi, Fahradin
Rémondeau, Paul
Nicolet-dit-Félix, Kléber
Bonnet-Eymard, Bénédicte
Dalmazzone, Didier
Hessler-Wyser, Aïcha
Ballif, Christophe
DOI
Abstract
The increasing demand for renewable energy sources has driven significant advancements in photovoltaic (PV) systems. The performance of these systems is, however, very sensitive to partial shading conditions which is important for Building Integrated PV (BIPV). Models that estimate the impact of partial shading on PV systems do not offer resolution at the cell level, tending to overestimate near shading losses. For example, the PVSYST approach simulates shading at the submodule level, assuming that if any part of a submodule is shaded, the entire submodule receives only diffuse irradiance. This neglects cases where only part of a cell is shaded, resulting in an overestimation of shading losses because it doesn't account for the partial shading of individual cells. This level of detail would improve the accuracy of near shading simulations used to design PV systems for partial shading conditions. This project develops two near shading models with cell level resolution using two different approaches, vertex projection and ray tracing. The ray tracing method considers light interactions with the system and surroundings, providing even more precision than geometric vertex projection. Both methods were validated on images of shadows observed on an outdoor module in Neuchâtel, Switzerland. Despite the increased computational complexity required for the higher resolution, optimization helped to reduce the simulation runtime by a factor of more than 20. The near shading simulation generates cell irradiance values, which will be used in a SPICE two-diode electrical model to simulate the I-V curve. This electrical simulation is particularly interesting for BIPV applications as it allows for a detailed analysis of the electrical performance under various shading conditions, crucial for optimizing energy yield in real-world installations. The simulation will be validated on measured outdoor I-V curves for the same experimental setup used to validate the shading model. The accuracy of the simulated electrical performance and total simulation runtime will be evaluated in several shading scenarios, from lightly shaded to fully shaded cells. Additionally, the shading model resolution, or number of points sampled per cell, will be varied to study the accuracy and runtime as a function of sampling resolution. Building on this work, we plan to further improve the ray tracing approach in spring 2025 by leveraging faster GPU simulations.
Publication Reference
EUPVSEC 2025, Bilbao, Spain
Year
2025