Dynamic Degradation Quantification of Wind Turbine High Speed Shaft Bearing Based on Oscillation Based Sparsity Indices

Abstract

Wind turbine (WT) high speed shaft (HSS) bearing fault is important due to its significant number of failures. However, due to its non-stationary operation under varying speed condition, it is a challenging issue to quantify the degradation of WT-HSS bearing by conventional bearing indices. Aiming at the aforementioned issue, in this paper, tunable Q factor wavelet transform (TQWT) preprocessed sparsity indices are proposed to achieve the dynamic degradation quantification of WT-HSS bearing. Firstly, based on the ability of TQWT in dynamic extraction of fault component of a bearing signal under varying speed condition, low oscillatory transient signal component is separated from a noisy commercial WT-HSS bearing signal by continuously adjustable tunable Q factor wavelet transform (TQWT). Then, considering the suitability of sparsity indices in quantifying extracted fault component based on its energy concentration, four representative sparsity indices namely kurtosis, gini index, negative entropy and reciprocal smoothness index are used to quantify the separated low oscillatory signal component as a measure of WT-HSS bearing health. The proposed indices show a better performance than original sparsity indices and an improved version of sparsity indices based on adaptive weighted signal preprocessing in dynamic degradation quantification of a commercial WT-HSS bearing.