Pvlib python is a community-supported open source tool that provides a set of functions and classes for simulating the performance of photovoltaic energy systems. pvlib python aims to provide reference implementations of models relevant to solar energy, including for example algorithms for solar position, clear sky irradiance, irradiance transposition, DC power, and DC-to-AC power conversion. pvlib python is an important component of a growing ecosystem of open source tools for solar energy (William F. Holmgren, Hansen, Stein, & Mikofski, 2018). pvlib python is developed on GitHub by contributors from academia, national laboratories, and private industry. pvlib python is released with a BSD 3-clause license allowing permissive use with attribution. pvlib python is extensively tested for functional and algorithm consistency. Continuous integration services check each pull request on multiple platforms and Python versions. The pvlib python API is thoroughly documented and detailed tutorials are provided for many features. The documentation includes help for installation and guidelines for contributions. The documentation is hosted at readthedocs.org as of this writing. A Google group and StackOverflow tag provide venues for user discussion and help.

Pvlib python is a community-developed, open-source software toolbox for simulating the performance of solar photovoltaic (PV) energy components and systems. It provides reference implementations of over 100 empirical and physics-based models from the peer-reviewed scientific literature, including solar position algorithms, irradiance models, thermal models, and PV electrical models. In addition to individual low-level model implementations, pvlib python provides high-level workflows that chain these models together like building blocks to form complete "weather-to-power" photovoltaic system models. It also provides functions to fetch and import a wide variety of weather datasets useful for PV modeling. pvlib python has been developed since 2013 and follows modern best practices for open-source python software, with comprehensive automated testing, standards-based packaging, and semantic versioning. Its source code is developed openly on GitHub and releases are distributed via the Python Package Index (PyPI) and the conda-forge repository. pvlib python's source code is made freely available under the permissive BSD-3 license. Here we (the project's core developers) present an update on pvlib python, describing capability and community development since our 2018 publication (Holmgren, Hansen, & Mikofski, 2018).

In this paper, we present the effect of installation parameters (tilt angle, height above ground, and albedo) on the bifacial gain and energy yield of three south-facing photovoltaic (PV) system configurations: a single module, a row of five modules, and five rows of five modules utilizing RADIANCE-based ray tracing model. We show that height and albedo have a direct impact on the performance of bifacial systems. However, the impact of the tilt angle is more complicated. Seasonal optimum tilt angles are dependent on parameters such as height, albedo, size of the system, weather conditions, and time of the year. For a single bifacial module installed in Albuquerque, NM, USA (35 °N) with a reasonable clearance (∼1 m) from the ground, the seasonal optimum tilt angle is lowest (∼5°) for the summer solstice and highest (∼65°) for the winter solstice. For larger systems, seasonal optimum tilt angles are usually higher and can be up to 20° greater than that for a single module system. Annual simulations also indicate that for larger fixed-tilt systems installed on a highly reflective ground (such as snow or a white roofing material with an albedo of ∼81%), the optimum tilt angle is higher than the optimum angle of the smaller size systems. We also show that modules in larger scale systems generate lower energy due to horizon blocking and large shadowing area cast by the modules on the ground. For albedo of 21%, the center module in a large array generates up to 7% less energy than a single bifacial module. To validate our model, we utilize measured data from Sandia National Laboratories’ fixed-tilt bifacial PV testbed and compare it with our simulations.

Real-time state estimation using a digital twin can overcome the lack of in-field measurements inside an electric feeder to optimize grid services provided by distributed energy resources (DERs). Optimal reactive power control of DERs can be used to mitigate distribution system voltage violations caused by increased penetrations of photovoltaic (PV) systems. In this work, a new technology called the Programmable Distribution Resource Open Management Optimization System (ProDROMOS) issued optimized DER reactive power setpoints based-on results from a particle swarm optimization (PSO) algorithm wrapped around OpenDSS time-series feeder simulations. This paper demonstrates the use of the ProDROMOS in a RT simulated environment using a power hardware-in-the-loop PV inverter and in a field demonstration, using a 678 kW PV system in Grafton (MA, USA). The primary contribution of the work is demonstrating a RT digital twin effectively provides state estimation pseudo-measurements that can be used to optimize DER operations for distribution voltage regulation.

Literature describes various methods for determining a series resistance for a photovoltaic device from measured IV curves. We investigate use of these techniques to estimate the series resistance parameter for a single diode equivalent circuit model. With simulated IV curves we demonstrate that the series resistance values obtained by these techniques differ systematically from the known series resistance parameter values used to generate the curves, indicating that these methods are not suitable for determining the series resistance parameter for the single diode model equation. We present an alternative method to determine the series resistance parameter jointly with the other parameters for the single diode model equation, and demonstrate the accuracy and reliability of this technique in the presence of measurement errors.