Research On Structural Statistical Energy Analysis And Fatigue Life Prediction Method

Today,with the development of high speed,high temperature,large and complicated of mechanical devices,the service environment becomes even more worse and the random factors increase greatly.As a result,more and more attention has been attracted to high-frequency vibro-acoustic coupling problems and structural fatigue damage problems in engineering,for instance,the aircraft subject to the vibro-acoustic environment which comes from propulsion systems and aerodynamic noise in flight can lead to the vibration of instrument and equipment,resulting in fatigue failure.When dealing with high-frequency problems,traditional deterministic analysis methods,Boundary Element Method(BEM)and Finite Element Method(FEM),for instance,have some limitation.Therefore,Statistical Energy Analysis(SEA)is developed to solve the problem and is widely used in engineering field.Frequency-domain fatigue life prediction method base on stress power spectral density(PSD)of structure is of simple thinking,of which the stress power spectral density method with zero-order-moments(SPSD0)can predict the fatigue life only based on the zero-order-moments of stress PSD is simpler and has higher computational efficiency.Time-domain fatigue life prediction method gets statistical information of stress amplitude and peak from loading history of structure to predict the fatigue life,of which the rain flow counting method's result is considered to be the most accurate fatigue life prediction method.But it has some limitation in engineering applications for its low computational efficiency.The dissertation focus on the high-frequency vibro-acoustic problem of an aircraft from a practical project based on SEA and the fatigue life prediction of typical stress spectral densities based on frequency-domain and time-domain methods,as follows:(1)Analysis of influence factors to high-frequency vibro-acoustic responseThe basic theory of statistical energy analysis method and the determination of the related statistical energy analysis parameters are introduced firstly,then the statistical energy analysis model of an aircraft is modelled,and numerical simulation method is used to discuss the effect of damping loss factor,absorption coefficient and the thickness of plate and shell on high-frequency vibro-acoustic response.The results show that:the changes of damping loss factor and the thickness of plate and shell have a great effect on cabin's pressure level and structure's acceleration response,while the changes of absorption coefficient has a great effect on cabin's pressure level but little effect on the acceleration of the structure.(2)Analysis of fatigue life prediction for unimodal stress spectral densityThe unimodal parabolic stress spectral density model is established,the fatigue damages under single band and wide band of unimodal stress process are calculated by SPSD0 and classical frequency-domain fatigue life prediction methods through a large number of numerical simulations,the damages result from different methods are also analyzed.The results show that:for narrow band process or wide band process that was divided into appropriate segments,the fatigue damage is not sensitive to exact curve of unimodal stress spectral density.(3)The analysis of fatigue life prediction for bimodal stress spectral densityThe triangular model of bimodal stress spectral density which is commonly observed in dynamic response of system is established,and the fatigue damages under single band and wide band of bimodal stress process are calculated by SPSD0 and classical frequency-domain fatigue life prediction methods by means of numerical simulation,the damages result from different methods are also analyzed.The simulation results show that:for narrow band process or wide band process that was divided into appropriate segments,the fatigue damage is not sensitive to the real curve shape of bimodal stress spectral density.The comparison of prediction results calculated from frequency-domain methods and rain flow counting method for bimodal stress spectral density with different characteristics further verify the correctness of the conclusion.