Precise backside irradiance modeling for bifacial PV is often unnecessary

A European research group has developed a new "empirical" method for measuring the backside irradiance of bifacial PV system. The proposed approach was tested across several European locations and it was found to enable annual performance calculations with a fixed backside irradiance share value.

Scientists from Slovenia’s University of Ljubljana and Spain’s University of Jaén have developed a novel empirical model for calculating the backside irradiance of bifacial PV systems.

“The core of the work is a new empirical model for backside irradiance,” corresponding author Kristijan Brecl told pv magazine. “Until now, all backside irradiance models were based on the geometry of the PV system and solar radiation values. Thus, to our knowledge, this is the first empirical model on the market. The results might not be extraordinary, but are comparable to the exact physical models.”

The model uses a Gaussian function to describe how the backside irradiance share changes with the solar azimuth angle.

The team explained that most backside irradiance models are based on analytical methods. They are based on physics and geometry, using site-specific data such as PV system geometry, albedo, and surroundings. They are considered to be more accurate, but cannot be easily applied to another site without knowing all the relevant parameters. An empirical model, on the other hand, uses measured data and statistical analysis. They are more versatile, but offer less accuracy on an hourly or daily basis. For the annual simulation, the team emphasized, they do provide acceptable results.

The proposed new empirical model was developed using nine bifacial PV modules that were installed in southern Spain. Five of them had their front side covered with a white film, and were used as backside irradiance sensors. Additional instrumentation included a plane-of-array pyranometer, a reference cell, and a nearby albedo meter.

The model simulates the backside irradiance by determining its share relative to the plane-of-array irradiance.

“The share of the back irradiance in the total irradiance received by a bifacial module is generally consistent throughout the day, with deviations in the morning and evening hours in summer when the sun shines from behind,” the group said. “Our new empirical model simulates the backside irradiance share as a ratio of the back-to-front irradiance with respect to the solar azimuth. The share is modelled by an exponential equation of inverse Gaussian distribution shape with the parameters derived from the measured data in the training period.”

The validation of the model was performed on PV modules at two other locations in Europe, namely Ljubljana, Slovenia, and Neuchâtel, Switzerland. In addition, the group has also compared the results of the empirical model to their own previously developed performance assessment model – the direct-diffuse power rating model for bifacial PV modules (DDPRbifi) model.

“The results will be used to improve the performance assessment model. Our aim in the future is to validate the model on big PV systems and for special locations/conditions,” Brecl added. “The main message of our current paper is that a detailed backside irradiance modelling is not essential for PV system performance assessment studies. However, an annual backside irradiance share can be used for fast and simple calculations.”

The new methodology was presented in “Is an exact backside irradiance modelling essential for bifacial PV systems?” published in Renewable Energy.