Novel early warning fault detection for wind-turbine-based DG systems

This paper will study condition monitoring signals of a distributed generation (DG) system not only due to the mechanical and electrical faults inside the wind turbines but also due to the grid system fluctuations. A novel feature extraction and characterisation method based on singularity detection of the monitoring data will be presented, aiming to identify the abnormal events and fault conditions as early as possible. The algorithm used to calculate Lipschitz values is given in the paper and efficient processing and storage of monitoring data is also discussed. The preliminary research has produced promising results.