Researchers at the Technical University of Madrid have conducted a comprehensive review of modeling tools used to measure the incidence angle modifier (IAM) – a factor describing how the optical efficiency of photovoltaic (PV) modules changes as the sun’s rays deviate from a line perpendicular to the module surface.

The team explained that IAM losses in PV systems are primarily reflection losses caused by changes in the optical properties of materials as light passes from air to the solar cells. These losses are more pronounced in panels affected by soiling. “Annual IAM losses, usually referred to as annual angular losses (AAL), account for between 1% and 5% of the yearly energy yield in utility-scale PV plants,” the researchers emphasized. “This is significant enough to be considered when modeling system performance.”

The research team described six IAM models: the ASHRAE model, the simplest approach, adopted by the American Society of Heating, Refrigerating, and Air-Conditioning Engineers (ASHRAE); the Air-Glass model, which considers only the air–glass interface and uses a single refractive index (n) value; the Martín-Ruiz model, specifically recommended in the IEC 61853-2 standard for describing corresponding IAM experimental results; the Sandia IAM database, which the scientists described as “no longer supported” and “deprecated”; the Physical model, based on Snell’s and Bouguer’s laws; and the Eye-Sensitivity model, which simulates how the human eye adapts to varying light levels, particularly in high-dynamic-range environments.

After evaluation, the team discarded the Sandia model for being unsupported and the Physical model for its complexity. The remaining four models were implemented to simulate front and rear faces of PV modules in SISIFO, a PV performance simulation developed by the Technical University of Madrid itself. “This freely accessible tool facilitates simulation and helps to reduce uncertainty in the estimation of final yield,” the research's corresponding author, Felipe Javier Ríos Ledesma, told pv magazine.

“This was followed by an extensive simulation exercise, covering both static and tracked PV arrays at three different latitudes, aimed at understanding the impact of IAM losses and the compatibility between the different IAM models,” they explained. “This allows software improvements to be validated against real PV plants, reducing uncertainty in annual energy calculations.”

To assess the real-world impact of IAM losses and the effectiveness of the proposed models, the researchers conducted simulations of front AAL radiation losses at three sites across different latitudes, using two types of monofacial PV installations based on fixed-structure and single-axis trackers, respectively.

The results indicated that the Air-Glass model overestimates annual angular losses for direct and reflected irradiance while underestimating losses for isotropic diffuse irradiance. Meanwhile, the Eye-Sensitivity model consistently predicted significantly lower AAL than the other models.

The Martín-Ruiz model stood out as the only model capable of calculating additional annual angular loss caused by soiling. “As a reference, for a 2% loss in perpendicular irradiance due to soiling, this model estimates an additional AAL of 0.5%, 0.9%, and 1.2% for single-axis trackers in Chile, Spain, and Sweden, respectively, and 1.0%, 1.3%, and 1.7% for static PV installations. This underscores our recommendation to incorporate this additional loss into other models that do not account for it,” the team noted.

The researchers concluded that their review provides a valuable tool for measuring IAM losses in PV modules with bare-glass or anti-reflective (AR) coatings. “Annual losses across models were consistent within 0.2%, ensuring reliable accuracy in all cases. Overall, annual energy losses range from 1% to 5%, depending on the IAM model used,” they stated.

“What caught our attention most was the notable discrepancy in energy yield predictions, where annual angular losses (AAL) can vary between 1 and 5% solely due to the choice of IAM model,” Ledesma emphasized. “This finding highlights a source of uncertainty that is often overlooked in bankable energy reports and photovoltaic plant performance assessments. Taking this effect into account is essential for improving the accuracy of economic models and, consequently, obtaining more realistic estimates of the return on investment in industrial projects.”

The research work is presented in “Angular losses in photovoltaic energy estimation: A review of models and a comparative analysis using the enhanced SISIFO simulation tool,” published in Renewable and Sustainable Energy Reviews.

Popular content

Ireland surpasses 2 GW of installed solar capacity

14 November 2025 Ireland's national solar capacity has risen to 2.1 GW as buildout has accelerated since the first utility-scale project came online in April 2022.