Research On Structural Health Monitoring Method For Fatigue Damage Of Carbon Fiber Composites Based On Active Lamb Waves

Carbon fiber reinforced polymer(CFRP)is widely used in aerospace,energy construction,national defense and military industry due to its high strength to weight ratio,high stiffness to weight ratio,good corrosion resistance,strong designability,fatigue resistance and light weight.However,in the process of material processing,component manufacturing and service,CFRP will continue to produce various kinds of defects and damages,among which fatigue damage is the most dangerous and the most important form of structural damage,accounting for 50%-90%of the total failure.Therefore,it is very important to discover the location and development of the damage in time.Based on active Lamb wave online monitoring technology,this paper carried out research on CFRP fatigue damage.The main research work in this paper was as follows:(1)A new damage factor was proposed to improve the damage probability-based diagnostic imaging method.Aiming at the problems of high misjudgment rate of damage location,low definition of damage imaging and poor visualization effect of existing probability-based diagnostic imaging methods,a fatigue damage probability-based diagnostic imaging method of carbon fiber composites based on time of flight(TOF)damage factor was proposed.Firstly,the time-of-flight characteristics of health reference signal and damage scattering signal of each excitation sensing channel were extracted by Hilbert transform;secondly,a new damage factor,TOF damage factor,was constructed based on the extracted features;finally,the damage probability-based diagnostic imaging method was improved by using the proposed damage factor,and the fatigue damage imaging experiments of carbon fiber composite plate under different fatigue load cycles were studied.The experiment result shows that compared with the existing probability-based diagnostic imaging method,the damage location error of this method is reduced by49.85%or more.(2)Aiming at the problem that the structural health monitoring method based on active Lamb wave will be disturbed by strong noise such as structural vibration and service environment in practical application,which makes the effective signal feature extraction inaccurate and affects the accuracy of damage location,a fatigue damage probability-based diagnostic imaging method of CFRP based on improved damage factor under strong noise background was proposed.In this paper,five different smoothing algorithms were used to smooth the signal,the smoothing effect of these five different methods was compared,and the best one was selected for research obtain more accurate time-of-flight feature under strong noise,and then To F damage factor was improved.Combined with the damage probability-based diagnostic imaging method,the damage under strong noise environment was analyzed.The experimental results showed that the proposed method can effectively locate the internal fatigue damage of structure in strong noise environment,and avoid the problem of inaccurate feature extraction and damage location in strong noise environment.The experiment result shows that compared with the existing probability-based diagnostic imaging method,the damage location error of this method is reduced by 63.7%or more.(3)Aiming at the problems that large prediction error and poor stability of generalized regression neural network(GRNN)model for the fatigue damage development of CFRP.Particle swarm optimization was used to optimize the GRNN model.Combined with the signal characteristics of cross-correlation damage factors,the fatigue damage prediction model of CFRP based on PSO-GRNN was constructed.The performance of the model was evaluated by four indexes:Autocorrelation Coefficient(R~2),Root Mean Square Error(RMSE),Mean Square Relative Error(MSRE)and Mean Absolute Error(MAE).The experimental results showed that the error of the model based on PSO GRNN is small,high prediction accuracy,the autocorrelation coefficient is up to 95.5%,and the stability is good.The prediction results did not change greatly with the change of model parameters.The error of GRNN model is large and easy to be affected by the change of parameters.The autocorrelation coefficient is low.In this paper,active Lamb waves were used to monitor the internal fatigue damage of CFRP.The experiment result shows that the research of this paper provides a new research idea for the internal fatigue damage location analysis,quantitative analysis,damage location analysis under strong noise environment and damage development predict of CFRP.