Wind Turbine Fault Detection based on Artificial Neural Network Analysis of SCADA Data
Jürgen Herp, Esmaeil S. Nadimi · University of Southern Denmark Research Portal (University of Southern Denmark) · 2015
Slowly developing faults in wind turbine can, when not detected and fixed on time, cause severe damage and downtime. We are proposing a fault detection method based on Artificial Neural Networks (ANN) and the recordings from Supervisory Control and Data Acquisition (SCADA) systems installed in wind farms. We establish a model for the normal behaviour of a wind turbine from considered fault-free data and test the proposed model on further data. We show that ANN can be used for early fault detection in wind turbines monitoring. Concerning vibrational levels in x and y directions we base our fault detection upon a generalized-likelihood-test. An upper and a lower control bounds are established for x and y respectively, given a minimum false alarm probability η based on the statistical characteristics of the data.