Impacto do carregamento das pás no desempenho de aerogeradores em Parque Eólico do Rio Grande do Norte

Abstract

The increasing demand for renewable energy has driven the expansion of wind power generation worldwide. However, the efficiency of wind turbines depends on multiple operational factors, including blade loading. This study analyzes the impact of blade loading anomalies on the performance of wind turbines in a wind farm located in Rio Grande do Norte, Brazil. The objective is to quantify energy losses resulting from blade load limitations and propose solutions to mitigate these effects. The research adopts a case study approach, using historical data from 2023 and 2024, collected through the Google Cloud Platform (GCP). The dataset includes operational variables such as generated power, wind speed, and blade load indicators. Data analysis was performed using statistical techniques and machine learning models, particularly Autoencoders, to detect anomalies in blade load values. The study identified eight wind turbines with significant limitations, three of which were directly related to blade load anomalies. These issues resulted in an estimated energy loss of 1,636.38 MWh, leading to financial losses of approximately R$490,913. The results emphasize the importance of continuous monitoring and advanced predictive maintenance strategies to optimize turbine performance and reduce operational losses. Proposed improvements include the enhancement of anomaly detection models, the development of interactive dashboards for real-time monitoring, and the adoption of predictive maintenance practices. These measures aim to increase wind turbine availability and efficiency, ensuring greater energy generation and economic viability for wind farms.