This project introduces a fault detection and isolation (FDI) system for wind farms based on SCADA (Supervisory Control and Data Acquisition) data. Instead of the traditional approach of monitoring each turbine individually, this project proposes a model-based FDI system that exploits correlations between measurements associated to neighboring turbines in order to detect and localize possible faults. This feature could enhance the performance of the FDI system, in terms of its capability for early detection of faults and the reduction of false detection and missed detection occurrences.
The main challenge is to account for the difference in wind conditions that each turbine is subject to, which can significantly modify the relative turbine behaviors. To this end, linear parameter-varying 2(LPV) models will be used to represent the relative dynamics between the turbines as a function of wind speed and wind direction, notably. These LPV models will be identified in the frequency domain since the modeling can be done in a user defined frequency band. Also, the continuous-time framework will be chosen because models related with the physics of the system can facilitate the design of the FDI system.
The project will be carried out as follows. First, a LPV model for monitoring a single wind turbine will be developed with focus on the pitch system. Second, the results obtained are extended for the LPV modelling of wind farms with focus on the data associated to turbine efficiency and overheating. Third, the obtained models are analyzed and a FDI system for wind farms is designed. This brings additional challenges to be addressed: how to best construct the LPV models for enhancing fault detection while managing complexity? How to automatically extract relevant data for the modeling? How to ensure a systematic design and tuning of the FDI system?
Finally, the validation on a wind farm simulator and with SCADA data from an on-shore wind farm made of ten 2.5 MW turbines will be performed.