Performance Modeling Tools Overview

Performance modeling tools predict the energy a user-defined PV system will produce in a given location, assuming it is properly installed with the components specified. Modeling tools use a series of mathematical equations to simulate power input and output values for specific electrical components, and then estimate PV system energy production on an hourly basis and aggregate these values to predict annual energy production. Industry stakeholders rely on performance model predictions for a wide variety of purposes: optimizing designs; informing proposals, contracts and performance guarantees; securing financing; calculating capacity factor; studying utility system impacts; and executing energy tests. System designers, integrators, project developers and researchers have access to both free and fee-based performance modeling tools to meet these needs.

Free Tools

The National Renewable Energy Laboratory (NREL) offers two free grid-connected PV system performance-modeling tools, with somewhat different target markets—PVWatts, for residential system integrators or preliminary performance estimates on larger projects; and the System Advisor Model, for advanced user groups such as project developers and equipment manufacturers.

PVWatts. NREL released the PVWatts Calculator ( in 1999, billing it as “a performance calculator for grid-connected PV systems.” The cloud-based software platform is one of NREL’s most trafficked websites, attracting more than 20,000 users per month. The authors of the NREL report “Validation of Multiple Tools for Flat Plate Photovoltaic Modeling Against Measured Data” refer to PVWatts as “a relatively simple PV performance estimation tool [with few inputs] designed to give users a starting point for evaluating the feasibility of a PV system.” As such, users primarily rely on PVWatts for estimating residential PV system performance or generating preliminary energy estimates for larger systems.

In September 2014, NREL released a new version of PVWatts that features an improved user interface and more-accurate algorithms. Whereas earlier versions of PVWatts assumed a default dc derate value of 77%, which notoriously underestimated system performance, the new version uses a default system loss percentage of 14% and generates 7–9% higher production estimates for fixed-tilt PV systems. The new version of PVWatts also allows users to select from different weather data, use a Google Maps–based roof layout tool, and specify advanced parameters such as dc-to-ac ratio and ground coverage ratio.

Based on the specified ground coverage ratio, PVWatts now incorporates basic estimates of interrow shading losses for single-axis tracker systems with or without backtracking. However, PVWatts relies on single, static annual percentage loss factors to estimate the impacts of shading, soiling and snow on production. Efficiency is the only module or inverter performance characteristic that users can define in PVWatts. By comparison, all of the other tools discussed in this article either include extensive inverter and module databases or allow users to import or manually input component-specific performance characteristics.

System Advisor Model (SAM) NREL developed SAM ( in 2005. It revises the desktop software program on an annual basis and released the most recent revision in June 2015. Like many of the other advanced modeling tools, SAM calculates PV system performance using mathematical submodels developed by governmental and academic organizations, including NREL, Sandia National Laboratories, the University of Wisconsin and others. For example, Sandia National Laboratories provides submodels used to define array and inverter performance. Richard Perez, a research professor at the University at Albany, State University of New York, created a widely used submodel for the complex process of translating horizontal surface irradiance values to plane-of-array irradiance values. The target market for SAM includes advanced user groups such as project developers, policymakers, equipment manufacturers and researchers.

When generating performance predictions, SAM users can choose between the simple PVWatts model and a more granular model with user-selectable submodels. The latter option lets users model system production based on module- or inverter-specific performance characteristics, as well as account for losses due to environmental variables such as soiling or snow. The platform includes powerful statistical analysis tools that allow users to run Monte Carlo simulations or study weather variability via a P50/P90 analysis. SAM also has more in-depth financial analysis capabilities than many other modeling tools. Users can choose between eight different financial models for PV systems, ranging from commercial power purchase agreements to third-party leases.

Article Discussion

Related Articles