Adapting the Durisch model for bifacial silicon and high concentration multijunction photovoltaic modules efficiency prediction

Efficiency models are valuable tools for predicting the performance of photovoltaic (PV) systems. By providing expected efficiency based on operating conditions, they enable comparison with actual measured efficiency. Significant discrepancies between predicted and measured values can indicate operational issues, such as measurement errors or equipment defects. This study focuses on the Durisch model, a widely used and consolidated efficiency model. We adapted the model to describe the efficiency of bifacial and high concentration multijunction devices (HCPV), incorporating different components of solar radiation. Our approach was validated using an extensive experimental setup with four distinct PV module types. Data filtering methods are discussed, and the normalized root mean square error (nRMSE) for the model with less restrictive filtering was 3.43 %, 2.42 %, 3.33 %, and 7.32 % for polycrystalline, monocrystalline, bifacial, and HCPV PV arrays, respectively. Different factors contribute to a relatively higher error for the HCPV array efficiency modeling, such as: the HCPV array's response to the beam irradiance only; the 90-m distance between the beam irradiance meter and the HCPV array; the HCPV array's sensitiveness to errors on tracking due to the use of concentrators; and the beam irradiance's changing more rapidly than the global irradiance.