Artistic BIPV model links 1,296 cells to power and hot-spot risk

A South Korean model evaluates 1,296 individual photovoltaic cells in each 36 × 36 colored building-integrated PV mosaic, tying an artwork’s pixel-level color to electrical output and hot-spot risk.

Researchers from Seoul National University of Science and Technology and Gachon University developed the work, which appeared in Energy as “Prediction Model for Power Output and Hot-spot Risk in Colored Photovoltaic Arrays with Artistic Image.”

Corresponding author Hyung-Jun Song told pv magazine that the novelty lies in directly linking a candidate image’s visual characteristics with the electrical performance and reliability of the entire series-connected array.

What does the model simulate?

The team modeled 50 masterpieces, including Johannes Vermeer’s Girl with a Pearl Earring and Vincent van Gogh’s Café Terrace at Night.

• Each artwork became a 36 × 36-pixel mosaic, with every one of the 1,296 pixels representing an individual series-connected PV cell.
• For each pixel, the team assigned the average color of its image region and estimated the minimum-loss, dual-band selective-reflection spectrum that most closely reproduced that color.
• The researchers calculated the resulting photocurrent from sunlight transmitted through the colored layer, accounting for the underlying cell’s wavelength-dependent internal quantum efficiency.

A single-diode equivalent circuit calibrated to commercial TOPCon cells supplied each cell’s current-voltage characteristics at 300 K. The team simulated the fully series-connected array in LTspice.

Song told pv magazine: “Unlike a uniformly colored module, an image-based array contains many differently colored cells. Because each color requires a different optical reflection profile, the photocurrent generated by each cell can vary across the array.”

The simulations compared configurations without bypass diodes against setups with one diode protecting between one and 36 cells. They assessed maximum power output and counted cells under reverse bias at the maximum power point as an indicator of potential hot-spot risk.

What changes across the artworks?

Song said: “One of the most surprising findings was that the overall brightness or colorfulness of an image does not necessarily determine its photovoltaic performance. Instead, the spatial distribution of brightness plays a much more important role.”

• Without bypass diodes, the Café Terrace at Night array produced 91.7% of the output of an equivalent black reference array.
• Girl with a Pearl Earring reached 74.3% of the black reference under the same configuration.

The study also found that even a small number of extremely bright pixels can limit the output of the whole series-connected array. Brightness distribution, rather than overall brightness or colorfulness alone, sets the more consequential electrical constraint.

Can bypass diodes solve the mismatch?

The results exposed a trade-off: more bypass diodes can improve power output, but they can also increase reverse-bias conditions and the associated hot-spot risk.

• For Girl with a Pearl Earring, one bypass diode per cell raised normalized output from 74.3% to 84.4% of the black reference, a 13.6% relative improvement.
• Yet 24 of the 1,296 cells, one in every 54, operated under reverse bias at the maximum power point.

Selective darkening of those 24 pixels raised normalized output to 84.8% without bypass diodes. Under the simulated conditions, the change also avoided reverse-biased cells.

What can designers screen before full simulation?

The team developed a framework that classifies artistic images by expected power performance and hot-spot risk using two image-derived parameters: the brightest pixel and overall brightness variation.

Researchers tested the framework with additional images, showing its potential to assess candidate artwork’s electrical suitability without requiring a full circuit-level simulation for every image. For BIPV designers, the result offers an earlier engineering check before engineers model every cell and bypass-diode configuration.

What happens next?

The team plans three follow-ups:

• Incorporate measured reflectance spectra into a more realistic optical model.
• Expand the image database.
• Develop an AI-assisted tool that assesses uploaded images and suggests modifications to improve output and reduce hot-spot risks while preserving their appearance.

The decisive next milestone is the move from idealized reflection spectra to measured reflectance, alongside broader image testing. If that work confirms the present simulations, the two-parameter screen could support a pre-simulation design check for BIPV manufacturers; for now, the 91.7%, 84.4% and 84.8% figures are modeled rather than operating-array results.