Solar input data for photovoltaic performance modeling

Abstract

One of the most critical inputs to a photovoltaic (PV) energy model is the solar data set, which establishes the site's irradiance and weather variability. For long-term energy estimates, the solar data set is expected to represent the long-term climatological conditions on-site. While modeled solar data sets are available, the quality of these data vary by data source as well as regionally. The result of using a poor quality solar input data set is higher uncertainty in the energy production estimated from the model; conversely, a more accurate solar input data set can improve the confidence in the energy production estimate. As the solar industry begins to recognize the value of increasing confidence in PV performance modeling predictions, an increased focus on quality input solar data for PV energy estimation models is expected. Publicly available data sources were evaluated with respect to their suitability as input data for PV energy estimation. These included modeled data sources, publicly, available reference station data, and site-specific measured data. The results of a research study conducted at nine locations throughout the United States show that both the magnitude and the distribution of input solar data sets affect energy. The value of on-site solar data collection and its ability to reduce uncertainty from between 2% to 5% is presented, as demonstrated from a case study from a site in the United States Desert Southwest.