Adaptive signal processing strategy for a wind farm system fault accommodation

In order to enhance the `sustainability' of offshore wind farms, thus skipping unplanned maintenance operations and costs, that can be important for offshore systems, the earlier management of faults represents the key point. Therefore, this work studies the development of an adaptive sustainable control scheme with application to a wind farm benchmark consisting of nine wind turbine systems. They are described via their nonlinear models, as well as the wind and wake effects among the wind turbines of the wind park. The fault tolerant (i.e., sustainable) control strategy uses the recursive estimation of the faults provided by nonlinear estimators designed via a nonlinear differential algebraic tool. These estimators are not affected by the model uncertainty and the wake effects among the wind turbines. This work exploits also a data-driven method used for estimating the analytical form of these disturbance functions, which are employed for obtaining the nonlinear fault reconstructors. Note that purely analytic approaches, where the model nonlinearity and the disturbance decoupling features are directly taken into account, may lead to more complex design tools. This aspect of the study, together with the more straightforward solution based on a data-driven scheme, is the issue when online applications are proposed for a viable implementation of the proposed solutions. The benchmark is exploited to verify the features of the developed strategies with respect to various fault situations and unavoidable model-reality mismatch.