The PV performance ratio paradox: annual data from large-scale, real-world PV systems show negligible meteorological and technical impact and points to dominant human factors

TL;DR

This study reveals that in large-scale PV systems in Brazil, human factors outweigh meteorological and technical variables in influencing performance ratio, suggesting a focus on installation and maintenance quality for optimization.

Contribution

It demonstrates that meteorological and technical factors have negligible impact on PV performance ratio, highlighting the importance of human factors in large-scale PV system efficiency.

Findings

Negligible effects of meteorological variables on PR

Human factors significantly influence PV performance

Peak PR value estimated at 78.85%

Abstract

Performance ratio (PR) is a established measure of efficiency of photovoltaic (PV) systems. While previous research demonstrated the effects of meteorological and technical variables on PR, a gap persists in the literature on which variables strongly influence PR in large-scale, real-world, heterogeneous PV systems. This paper aims to fill this gap, applying data-driven models to PV systems located in Rond\^onia State, Brazil, to identify which variables strongly influence annual PR, and, hence, should be the target for optimization. Surprisingly, only negligible effects were found between meteorological and technical variables on annual PR, indicating that human-factors (such as installation, monitoring, and maintenance quality) might have a stronger effect. These findings indicates that, to improve performance of PV systems, policy makers could focus on creating educational programs to teach PV installers and technicians how to properly install, monitor, and maintain modern PV systems. Through estimating the probability density functions of PR, its peak value was found as 78.85% (mean 77.52%, 95% confidence interval of 76.12% to 78.84%, and 95% prediction interval of 58.83% to 92.70%). A map of annual final yield was developed for Rond\^onia State and can be used by entrepreneurs to quickly and cheaply estimate energy production.