Validation of Bifacial Photovoltaic Simulation Software against Monitoring Data from Large-Scale Single-Axis Trackers and Fixed Tilt Systems in Denmark

Nicholas Riedel-Lyngskær,Djaber Berrian,Jan Vedde,Peter Behrensdorff Poulsen,Alexander Aguilar Protti,Daniel Alvarez Mira,Joris Libal

doi:10.3390/app10238487

Nicholas Riedel-Lyngskær, Djaber Berrian + Show 5 more

Open Access

https://doi.org/10.3390/app10238487

Copy DOI

Abstract

The size and number of utility-scale bifacial photovoltaic (PV) installations has proliferated in recent years but concerns over modeling accuracy remain. The aim of this work is to provide the PV community with a validation study of eight tools used to simulate bifacial PV performance. We simulate real 26 kilowatt-peak (kWp) bifacial arrays within a 420-kWp site located in northern Europe (55.6° N, 12.1° E). The substructures investigated include horizontal single-axis trackers (HSATs) and fixed tilt racks that have dimensions analogous to those found in utility-scale PV installations. Each bifacial system has a monofacial reference system with similar front side power. We use on-site solar radiation (global, diffuse, and beam) and albedo measurements from spectrally flat class A sensors as inputs to the simulation tools, and compare the modeled values to field measurements of string level power, rear and front plane of array irradiance, and module temperature. Our results show that state-of-the-art bifacial performance models add ~0.5% uncertainty to the PV modeling chain. For the site investigated, 2-D view factor fixed tilt simulations are within ±1% of the measured monthly bifacial gain. However, simulations of single-axis tracker systems are less accurate, wherein 2-D view factor and 3-D ray tracing are within approximately 2% and 1% of the measured bifacial gain, respectively.

Highlights

Bifacial photovoltaics (PV) has entered the mainstream market in recent years due to enhanced energy yields, which are enabled by the conversion of light impinging on the module’s backside into useable photocurrent
As more bifacial model validation studies are published from sites around the globe, the expectations that PV buyers and investors have regarding bifacial field performance will be in better alignment with actual field performance
When the measured—as opposed to calculated—angular position was used in the simulation, we found that the mean bias error (MBE) improved slightly from −1.1 to −0.7 W·m−2, but the mean absolute error (MAE), changed by less than 0.1 W·m−2

Summary

Introduction

Bifacial photovoltaics (PV) has entered the mainstream market in recent years due to enhanced energy yields, which are enabled by the conversion of light impinging on the module’s backside into useable photocurrent. Sci. 2020, 10, 8487 to the actual performance of fielded systems remains an ongoing task for the PV industry [2,3,4,5,6]. As more bifacial model validation studies are published from sites around the globe, the expectations that PV buyers and investors have regarding bifacial field performance will be in better alignment with actual field performance. If such validations of bifacial simulations are found to be within acceptable agreement, this has the potential to de-risk bifacial PV investments and lower soft costs. A key aim of this study is to contribute to this ongoing bifacial performance model validation effort

Objectives

Methods

Results

Conclusion