Enhancing Wind Turbine Performance: Statistical Detection of Sensor Faults Based on Improved Dynamic Independent Component Analysis

Efficient detection of sensor faults in wind turbines is essential to ensure the reliable operation and performance these renewable energy systems. This paper presents a novel semi-supervised data-based monitoring technique for fault using SCADA (supervisory control data acquisition) data. Unlike supervised methods, proposed approach does not require labeled data, making it cost-effective practical turbine monitoring. The builds upon Independent Component Analysis (ICA) approach, effectively capturing non-Gaussian features. Specifically, dynamic ICA (DICA) model employed account temporal dynamics dependencies observed signals affected by faults. process integrates indicators based on I2d, I2e, squared prediction error (SPE), enabling identification different types are combined with Double Exponential Weighted Moving Average (DEWMA) chart, known its superior detecting small magnitudes. Additionally, incorporates kernel density estimation establish nonparametric thresholds, increasing flexibility adaptability types. study considers various faults, including bias precision degradation freezing evaluation. results demonstrate that outperforms PCA traditional ICA-based methods. It achieves high rate, accurately identifying while reducing false alarms. could be promising proactive maintenance, optimizing reliability