Monitoring Research Of Fatigue Process Of Wind Turbine Blades Based On The Acoustic Emission

With the further worsening of the energy crisis and environmental damage,it has become a consensus to seek a kind of new renewable energy for the world.Wind energy,a renewable and clean,has been caused increasingly concern among people.Wind turbine blade is the key to obtaining wind energy with the wind turbine.Due to the limiting of conditions,however,the blades are prone to appear defects,such as debonding,delamination,fiber fracture and so on.The blades with such defects are often easy to fail under stress and cause great economic losses.In order to ensure the long-term reliable operation for the wind turbines,therefore,it is of great significance to know the damaging process of wind turbine blades well.Acoustic emission is a new non-destructing method,which can realize the dynamic monitoring of products in service.In these years,acoustic emission technology has obtained wide application in health monitoring of blades.Based on the research of wind turbine blades at home and abroad,the experiment was attempted to apply acoustic emission technology to the dynamic monitoring of fatigue process of wind turbine blade.So research about Lamb Source location;the pattern recognition of different damages of wind turbine blades and the monitoring of fatigue process of full scale wind turbine blades are studied.The veracity of the velocity is the key to locate the position of a fault using the acoustic emission technology in the Time Difference of Arrival.However,the frequency dispersion and modes generated from the propagation of the wave in a thin plate make it hard to acquire the wave speed,accurately.Based on the physical mechanism of the propagation of Lamb wave,the combination of complementary ensemble empirical mode decomposition is used combined with the theory of Lamb wave to analyze the signal under various excitations for locating the correctness of the acoustic emission source in the wind turbine blade by leading crack.The result shows that the kind of wave plays a main role in the plate when the force is parallel to the plate;another wave has the same role in the plate when the force is perpendicular to the plate,respectively.There is a higher velocity,high frequency about 150 kHz for the former,compared to the wave of lower frequency,and the later has the lower velocity,about frequency on 50 kHz.It is proved that using the former,with a lower attenuation and for amplitude,and without frequency dispersion to calculate acoustic emission source is better for wind turbine blade.It is difficult to extract for the fault feature of blade in service in the early period.Then,a new method of fault diagnosis is proposed for blade based on the variational mode decomposition(VMD)combined BP neural network.Firstly,a signal originated from blade was decomposed by VMD,and the intrinsic mode functions(IMF)containing main feature information were selected by variance contribution ratios.Then,the Energy Entropy of the IMF selected were calculated and as a feature vector for different defects.Finally,in order to verify the accuracy of the of feature vector selected,different entropy vectors energy of defect s were input into BP neural network to achieve a better result for pattern recognition.The results show that the recognition rate for fault was more than 90%,that the way of VMD combined energy entropy with BP neural network can realize defect recognition for blade in early period,and that be applied in practice.Finally,a fatigue experiment was performed on a full size wind power blade during the process of fatigue was monitored by a Non Destructive Testing of the acoustic emission.In order to expose fatigue damage mechanisms and the location of damage,power spectrum and the Complete Ensemble Empirical Mode Decomposition with Adaptive Noise(CEEMDAN)were used to the signal collected from different fatigue stages.The result shows that with the number of fatigue increasing,the failure of fatigue of blade appearances in certain position,and difference at different stages during process of fatigue.Some micro cracks of the matrix appeared on the blade,when the number of fatigue reached 1×105 times.When the number of fatigue was up to 5×105 times,the blade showed a slight crack growth in internal of blade,and the frequency are mainly around about 3 0 kHz;when the number of fatigue was around one million times,Fatigue appeared in surface of blade and the frequency shifted from 50kHz to 150 kHz;When the number of fatigue reached 1.5 million times,cracking for matrix,pulling out for fiber could be viewed.The main frequency focused on 60 kHz and 300kHz;When the number of fatigue reached 2 million times,the appearance went worsened,and matrix cracking and fiber fracture all are observed.The study not only provides a certain method for the dynamic monitoring of the blade,but also provides reference for the improvement of the blade structure.