Wind turbine blades fault detection based on principal component analysis

Abstract

This paper presents a new approach to detect faults in wind turbine blades. This approach is based on Principal Component Analysis (PCA) of the vibration signal. The residual matrix signals for healthy and faulty system were compared by applying the crest factor. It contains information extracted from the PCA and the faults were found from the comparisons. The experimental work was carried out using three bladed wind turbine. The cracks were simulated on the blade with diameters (3 mm, 6 mm, 9 mm and 12 mm), all had a consistent depth 3 mm. The tests were carried out for two rotation speeds; 250 and 360 rpm. The results showed that PCA of vibration based condition monitoring is a promising technique because it contains information on all the components of the wind turbine contained in the vibration signal. The crest factor was calculated for the PCA residual matrix. The novel approach successfully differentiated the signals from healthy system and system containing cracks in a turbine blade.