Estudo de caso do impacto da sujidade na eficiência energética de sistemas fotovoltaicos conectados à rede

Abstract

This paper presents a study and operational analysis of soiling effects on the energy efficiency of grid-connected distributed photovoltaic generation systems. The study is based on historical generation data obtained from solar companies in the state of Pará, Brazil, aiming to identify operational patterns associated with performance losses caused by dirt accumulation on photovoltaic modules. The proposed methodology applies generation normalization by peak sun hours, moving average filtering, and soiling ratio analysis, implemented in an interface developed with Python, Flask, and React, in which the user uploads the installation generation data to obtain the diagnosis and classification of soiling-related losses. A real validation case is presented using the L A Amaral installation, for which daily generation data and photographs of the modules before and after the cleaning procedure were collected. The results confirm the effectiveness of the approach: a severe deviation of-12,3% was identified in April 2025, and post-cleaning analysis revealed an 15.89% recovery in daily generation, rising from an average of 171,4 kWh to 185,7 kWh, with a calculated baseline of 45,32 kWh/HSP. The findings indicate that continuous operational monitoring can support preventive maintenance planning, reducing energy losses and financial impacts in distributed photovoltaic systems.