pvlib is the open reference implementation of photovoltaic performance models. This guide maps what it can do today, stage by stage, so you can find a model, a data source or a place to contribute.
Every pvlib model sits somewhere on this chain. Pick a step to filter the topic index and the API catalog below to just that step.
ModelChain runs the whole sequence for you; plain functions let you swap any single step.
The classes add no physics of their own. They wrap the same functions so you can choose how much bookkeeping pvlib does for you.
Call each model directly with pandas Series or NumPy arrays. Full control, nothing hidden.
# sun position, then POA irradiance
sp = pvlib.solarposition.get_solarposition(
times, 33.45, -111.98)
poa = pvlib.irradiance.get_total_irradiance(
30, 180, sp.apparent_zenith,
sp.azimuth, dni, ghi, dhi,
model="perez", dni_extra=dni_extra)
Keep site and hardware parameters on objects, then call methods that pass them along for you.
site = pvlib.location.Location(
33.45, -111.98, tz="US/Arizona")
sp = site.get_solarposition(times)
system = pvlib.pvsystem.PVSystem(
surface_tilt=30, surface_azimuth=180,
module_parameters=mod,
inverter_parameters=inv)
Infers suitable models from your system parameters and runs from weather, from POA or from effective irradiance.
mc = pvlib.modelchain.ModelChain(
system, site, aoi_model="physical")
mc.run_model(weather)
mc.results.ac # AC power
mc.results.cell_temperature
53 topics grouped into 11 domains, each with the named models it implements and where they live in the API. This is the seed of a contributor knowledge base.
Pulled from the API reference and docstrings on the main branch. Open a group to browse it, or search to open every group with a match.
Free sources for every topic: national lab reports, open-access papers, data provider documentation and pvlib's own guides. Each entry has a short summary in our own words and a link to the original. pvlib's documentation is included in full.
Loading the knowledge base…
The Energy Expert answers from this knowledge base and cites the references it used. Answers can still be wrong, so check the linked sources before relying on a number.
Try one of these:
pvlib fetches and parses irradiance and weather from 17 providers and networks, and reads the common file formats. Most functions return a DataFrame with pvlib's standard column names.
| Source | Kind | Access | Coverage | Functions |
|---|
v0.16.0 and v0.16.1 shipped on 26 September 2026. v0.16.2 is planned for December 2026 and is already collecting fixes.
About half of v0.16.2's notes so far are documentation fixes. Issues labelled good first issue and easy are the project's stated entry points.
Five areas have one example or fewer: decomposition, IV modelling, solar position, floating PV and system models.
New data sources arrive almost every release. Each needs a fetcher, a parser, variable mapping and remote-data tests.
The library grows by adding peer-reviewed models with references, tests against published values and a whatsnew entry.