No abstract

Wind turbine performance monitoring is a complex task because the power has a multivariate dependence on ambient conditions and working parameters. Furthermore, wind turbine nacelle anemometers are placed behind the rotor span and the control system estimates the upwind flow through a nacelle transfer function: this introduces a data quality issue. This study is devoted to the analysis of data-driven techniques for wind turbine performance control and monitoring: operation data of six 850 kW wind turbines sited in Italy have been employed. The objective of this study is an assessment of several easily implementable techniques and input variables selections for data-driven models whose target is the power of a wind turbine. Three model types are selected: one is linear (Principal Component Regression) and two are non-linear (Support Vector Regression with Gaussian Kernel and Feedforward Artificial Neural Network). The models' validation provides meaningful indications: the linear model in general has lower performance because it can not reproduce properly the non-linear pitch behavior when approaching rated power. Therefore, it is concluded that a non-linear model should be employed and the achieved mean absolute error is of the order of 1.3% of the rated power. Furthermore, the errors are kept at the order of 2% of the rated power for the models whose input is the rotor speed instead that wind speed: this observation supports that, in case it is needed because of nacelle anemometer biases, the power monitoring can be acceptably implemented using the rotor speed.

No abstract

No abstract

The importance of accurately forecasting the power production of wind farms is boosting the development of meteorological models and their processing. This work is a discussion of different forecast configurations for predicting the day ahead production of a wind farm sited in a moderately complex terrain. The numerical weather prediction (NWP) model MetCoOp Ensemble Prediction System with 2.5 km resolution focusing on the wind farm area is dynamically downscaled by the computational fluid model (CFD) model WindSim. The transfer of the NWP model to the CFD model can be done using NWP results from various heights above ground and using all or parts of the nodes of the NWP model within the wind farm area. In this work, many different forecasting configurations are validated and the impact on the forecast performance is discussed. The NWP-CFD downscaling results are compared to a day ahead forecast obtained through ANN methods and to the observed production. The main result of this work is that a deterministic downscaling method like CFD simulations can perform as good or better than statistical approaches when using high-resolution NWP models and more NWP model data.