Markovian jump system approach for the estimation and adaptive diagnosis of decreased power generation in wind farms

Abstract

In this study, a Markovian jump model of the power generation system of a wind turbine is proposed and the authors present a closed-loop model-based observer to estimate the faults related to energy losses. The observer is designed through an H ∞ -based optimisation problem that optimally fixes the trade-off between the observer fault sensitivity and robustness. The fault estimates are then used in data-based decision mechanisms for achieving fault detection and isolation. The performance of the strategy is then ameliorated in a wind farm (WF) level scheme that uses a bank of the aforementioned observers and decision mechanisms. Finally, the proposed approach is tested using a well-known benchmark in the context of WF fault diagnosis.