pv-system-model

This is a model for PV plants/systems based on pvlib [1]. One relatively novel feature is that it allows for row-to-row shade, using many concepts from [2], with the addition of cross-axis slope while still allowing for bifacial modules.

Features

Common modeling steps

The model_pv_power function in pv_model.py combines many common modeling steps, including:

• solar position
• surface tilt and azimuth for a number of tracking options (or fixed tilt):
  • standard backtracking
  • slope-aware backtracking
  • above options, modified with custom gcr settings (a single value or separate morning/afternoon values)
  • fully custom tilt/azimuth timeseries (calculated separately by the user and passed to the function)
• irradiance transposition with a selectable model (default of perez-driesse)
  • optionally, measured POA data can be used instead
• spectral modifier adjustment with pvlib.spectrum.spectral_factor_firstsolar if wanted
• incident angle modifier (iam) adjustment
• cell temperature calculation with pvlib.temperature.faiman
• dc power with pvlib.pvsystem.pvwatts_dc
  • including a user-defined dc loss fraction
• ac power with pvlib.inverter.pvwatts

Less common features

It also includes less common features that may be useful, including:

• separate rear-side irradiance with pvlib.bifacial.infinite_sheds.get_irradiance for bifacial system
  • infinite_sheds is only used for the rear to allow for cross-axis slope to be included in front-side irradiance calculations, as infinite_sheds currently only works with flat terrain.
• shade loss calculations for thin-film and common crystalline modules in portrait orientation
  • linear (thin-film) and non-linear (crystalline) shade losses
  • crystalline modules can be either older-style modules with square cells or modern "twin" modules with half-cut cells
  • allows for multiple "courses" of modules, e.g., "2P" where modules are in portrat two-high up the racking
  • Note: Crystalline modules in landscape are not an option currently. Anecdotally, there are vary few utility-scale fixed tilt solar projects in the US with crystalline modules in landscape orientation. I'm less certain about tracking projects with landscape crystalline, but those would likely use backtracking, in which case shade losses are not typically a focus area.

Inputs and outputs

The model_pv_power function takes in:

• a resource data DataFrame, similar to the format that is returned by pvlib.iotools.get* functions
• many plant/system specifications (which can also be passed in as a Python dictionary) that follow conventions and definitions in https://github.com/williamhobbs/pv-plant-specifications, specifically pv-plant-specification-rev6.csv

Users should make sure to select the appropriate input for shade_loss_model:

• 'non-linear_simple' for crystalline silicon modules with older-style cell stringing (typically square cells, three strings in series)
• 'non-linear_simple_twin_module' for crystalline silicon modules with modern stringing, two sets of parallel cell strings (typically half-cut cells)
• 'linear' for an module with a linear shade response (typical of thin-film like CdTe or CIGS)

The function returns:

• an ac power timeseries
• a modified version of the resource data DataFrame that includes the addition of modeled POA irradiance and modeled cell temperature. This might be useful for comparison with measured values from operating plants for performance engineering purposes.

Examples

See quick_demo.ipynb and https://github.com/williamhobbs/2025-pvrw-trackers for some example uses. It is also used in:

• https://github.com/williamhobbs/energy-forecasting-tools
• https://github.com/williamhobbs/PVPMC_2025
• https://github.com/williamhobbs/PVSC-2025-daily-energy-forecaster

References

[1] Anderson, K., Hansen, C., Holmgren, W., Jensen, A., Mikofski, M., and Driesse, A. “pvlib python: 2023 project update.” Journal of Open Source Software, 8(92), 5994, (2023). DOI: http://dx.doi.org/10.21105/joss.05994.

[2] Hobbs, W., Anderson, K., Mikofski, M., and Ghiz, M. "An approach to modeling linear and non-linear self-shading losses with pvlib." 2024 PV Performance Modeling Collaborative (PVPMC). https://github.com/williamhobbs/2024_pvpmc_self_shade