Aiming at the problem of ambiguous characteristics and poor discriminative properties of blade biofouling in marine current turbines(MCTs),a blade biofouling detection method based on adaptive proportional frequency sampling(APFS)was proposed.Firstly,the current signals of MCTs under differ-ent operating conditions were collected,and the instantaneous frequencies of the stator current signals were extracted by Hilbert transform;then,the instantaneous frequency sequence was sampled and the non-stable instantaneous rotation frequency was transformed into a stable value;finally,the sampling iter-ation stopping threshold was set by using the permutation entropy,and the re-sampled instantaneous fre-quency sequence was classified as a sample as a fault feature matrix to establish a principal component a-nalysis detection model for attachment detection.An experimental platform based on a 0.23 kW,MCT prototype was constructed to verify the effectiveness of the proposed method,and the experimental results show that the proposed algorithm has a low false alarm rate as low as 0.25%against the variable operat-ing conditions caused by the current speed change,and has high detection accuracy and robustness.
关键词
海流发电机/故障检测/叶片附着物/希尔伯特变换/自适应频率正比采样/主元成分分析
Key words
marine current turbines/fault detection/blade biofouling/Hilbert transform/adaptive pro-portional frequency sampling/principal component analysis
维普
万方数据