Research On Vibration Damageprediction Technology Ofpackage Based On Hilbertmarginal Spectrum And RVM

The rapid development of e-commerce has promoted the continues expansion of the logistics express industry.The packaging problem of the commodity transportation process has become the people’s attention.In order to save resources and improve the utilization of resources,packaging recycling is an effective way.During the transportation,the package suffers from the damage caused by the vibration.In order to ensure the safety of the goods,research on the damage level of the package and accurately predicting the future damage status of the package are the basis for judging whether the package can be recycled or not.In this paper,a damage prediction method based on the RVM model is proposed by analyzing the vibration signal characteristics of the package.The main research contents are as follows:Firstly,the damage mechanism of the package is analyzed from the perspective of fatigue cumulative damage theory and forced vibration dynamics analysis model.The accumulation of vibration damage can cause changes in the damping and stiffness of the package.The simulation analysis of the kinetic model also shows that the amplitude of the distribution changes at each frequency.Under the influence of external vibration environment and load,the damage of the package is gradually accumulated.The package is damaged when it exceeds the limit that the package can withstand.Thus,a fatigue vibration test method is proposed to provide support for the extraction of damage information.Secondly,the method of extracting vibration damage state features is studied.By analyzing and comparing different decomposition methods,the EMD decomposition method is more suitable for the vibration signal of package.The Hilbert transform process is performed on the decomposed signal to obtain the Hilbert marginal spectrum.On this basis,the three feature quantities of the marginal spectrum are constructed and applied to the multi-distances morphological similarity assessment method to calculate the damage index.Through the real experiment and the data analysis,the effectiveness of the feature extraction approach is demonstrated.Thirdly,an RVM-based package damage prediction model is constructed.To enhance the prediction accuracy of the traditional RVM model,an improved method of RVM is proposed.The Moth-Flame Optimization algorithm is applied to optimize the kernelparameters of the model,and then the damage index obtained by the previous analysis is used as the input of the model to predict the damage.The results show that the improved multi-core RVM model not only obtains a higher prediction accuracy but also is more suitable for damage prediction of the packages.Finally,the damage detection system of package based on Lab VIEW development platform is designed.By the test analysis,the effectiveness of the detection system is proved.