Michela Moschella, Alessandro Betti, Emanuele Crisostomi and Mauro Tucci.
The increasing penetration level of energy generation from renewable sources is demanding for more accurate and reliable forecasting tools to support classic power grid operations (e.g., unit commitment, electricity market clearing or maintenance planning). For this purpose, many physical models have been employed, and more recently many statistical or machine learning algorithms, and data-driven methods in general, are becoming subject of intense research. While generally the power research community focuses on power forecasting at the level of single plants, in a short future horizon of time, in this time we are interested in aggregated macro-area power generation (i.e., in a territory of size greater than 100000 km2) with a future horizon of interest up to 15 days ahead. Real data are used to validate the proposed forecasting methodology on a test set of several months.
A scalable predictive maintenance model for detecting wind turbine component failures based on SCADA data
Lorenzo Gigoni, Alessandro Betti, Emanuele Crisostomi and Mauro Tucci.
This work presents a novel Predictive Maintenance model for the main Wind Turbine components based on SCADA data and combination of Machine Learning and Classical Statistical approach. According to the impact of maintenance cost on the wind power Levelized Cost of Energy (LCOE), the components Gearbox, Generator, and Main Bearing have been modelled. The model was trained on historical nominal behavior periods of components specific SCADA tags. Test campaign was divided in two stages: test on historical faults for model training and validation, and Real-Time test for proper integration in the plant operators’ activities. Test on historical faults of six wind farms located in Italy and Romania, corresponding to an overall installed nominal power of 283 MW, demonstrate outstanding capabilities of anomaly prediction up to 2 months before device unscheduled downtime. Additionally, test on 12- months Real-Time phase confirm its ability to detect several anomalies, therefore allowing the operators to identify root case and schedule maintenance before reaching the catastrophic stage.