Research On Fault Diagnosis For The Key Components Of The Transmission System Of Offshore Wind Turbine Installation Crane

The offshore wind turbine installation industry of China started late,and the relevant monitoring,diagnostic and maintenance techniques are not complete.The offshore wind turbine crane always works in the harsh environment,the interval time of regular inspection and maintenance is unconfirmed.Therefore,the key components of the transmission system are prone to malfunction during operation.The transmission system is the core part of the offshore wind turbine crane,in which the bearings and gears are easily damaged in the harsh offshore environment.The operational status and health of key components directly determine the accuracy and safety of offshore wind turbine installation operations.Therefore,the fault diagnosis of the key components of offshore wind turbine crane system is very important for improving the safety and reliability of offshore wind turbine installation.In the existing research results of mechanical fault diagnosis,the research on the early weak fault detection for the transmission system key components of offshore wind turbine crane is very scarce;in the high humidity and platform shaking offshore environment,special band pitting and fretting defects would appear in bearings which are the key components of the transmission system,while there is no research on fault pattern recognition for special defects so far.Therefore,this paper chooses the key components bearings as the primary research objects and gears as the secondary research objects.The vibration signal analysis and fault diagnosis are carried out for the detection of early weak faults and the identification of special defects generated in harsh offshore environment.The main contents of the paper are as follows:(1)The research on the weak early fault feature extraction method for key components of the transmission system is carried out.The transmission system will produce a lot of noise and a variety of interference components in operation,the features of early weak faults are submerged and difficult to extract.Aimming at this situation,the traditional kurtosis indicator is improved according to the window telescopic sliding principle of wavelet transform;combining the advantages of kurtosis and wavelet transform,a fault feature extraction method based on continuous kurtosis and wavelet instantaneous energy feature fusion is proposed.The feature fusion method can effectively extract the weak early fault feature.(2)The research on the weak early fault signal enhancement detection method for the key components of the transmission system is carried out.In the vibration signal of the transmission system,the feature frequency of the fault often exists in the form modulation signal,its energy is far weaker than noise and other components,and thus it is difficult to detect.According to the stochastic resonance principle which utilizes noise energy to enhance weak signals,the traditional stochastic resonance system is improved,and an adaptive multi-stable cascade stochastic resonance method is proposed in combination with adaptive optimization algorithm.The proposed method can remedy the deficiencies of the traditional stochastic resonance method,improve the utilization of noise energy and achieve the enhanced detection for weak fault signals.(3)Aimming at the fault pattern recognition of the transmission system vibration signal,two fault diagnosis methods of multiscale feature extraction based on multiscale singular value and multiscale morphology are proposed.Through the multiscale feature extraction,the the complete features of complex signal can be obtained,and then the dimensions of nonlinear sample features can be reduced according to the manifold learning theory,the obtained low dimensional features are beneficial to identification.The multiscale singular value manifold method can extract the inherent dynamic characteristics of the signal samples from time-frequency distribution images and eliminate the interference of phase differences at the same time.The multiscale morphological manifold method can observe signal samples from different scales in time domain and construct the sample features of signals according to the correlation of the observation results in different scales.The two multiscale methods can effectively extract the signal features and improve the identification accuracy of fault patterns.Experimental verification shows that,the two fault diagnosis methods of multiscale feature extraction can effectively identify the band pitting and fretting defects of the transmission system which appear in the harsh offshore environment.(4)Aiming at the multifractal characteristics of the fault signal of the transmission system,a multifractal manifold fault diagnosis method based on detrended fluctuation analysis is proposed.In order to overcome the shortcomings of traditional detrended fluctuation analysis method in mechanical fault diagnosis,the more precise local fluctuation characteristics are obtained by improving the fitting precision of local trend;at the same time,the single fractal feature is extended to the multiple fractal domain,and the continuous generalized Hurst exponents are extracted as the sample features from the multifractal logarithmic curve,then the comprehensive sample features are obtained.The verification results of several fault experiments show that,the proposed method has higher fault pattern recognition accuracy,and its recognition performance have better adaptability and stability in different mechanical systems.(5)Aiming at the condition monitoring and fault diagnosis demand of the offshore wind turbine crane transmission system,the monitoring scheme and fault diagnosis system of the transmission system are preliminarily designed.The fault diagnosis system mainly includes signal sample analysis,feature extraction,enhanced detection,pattern recognition and other functions,the system and the functions are tested and verified by experimental data.