Study On Storage Reliability Analysis And Life Prediction Method Of Aerospace Relay Based On Degradation Data

The aerospace relay is a kind of military automatic control electronic component,which is widely used in the weapon with the need of long-term storage.Correctly predicting the storage life of aerospace relay plays an important role in ensuring the use of weapons in the normal life cycle,and it can also provide theoretical guidance and technical support for “reasonable definite longevity and scientific life extension” of weapons.At present,the traditional test method can not quickly obtain the failure data,and it is relatively single that the method for predicting the storage life of aerospace relays.In view of the above problems,this paper studies the accelerated degradation test method and storage life prediction method of aerospace relay based on the analysis of the storage reliability and the failure modes of aerospace relay.Firstly,the influencing factors of the aerospace relay in the long-term storage process is analyzed and several influencing factors that mainly affect the performance degradation of the aerospace relay is determined.By establishing the fault tree with the contact resistance anomaly as the top event,initially determined that the temperature is the main environmental stress affecting the storage reliability of aerospace relay.On this basis,the growth and diffusion process of corrosion film on the contact surface of aerospace relay is studied,and the influence of temperature for the corrosion film formation of the contact surface of aerospace relay is determined.It provided a basis for the design of the degradation test of aerospace relay.Secondly,the accelerated degradation test method of aerospace relay is designed by determining the contents of test type,test stress,test stress level,sample quantity,monitoring period and monitoring parameters.The aerospace relay degradation test is carried out based on the test system,obtained the experimental data such as contact resistance,pick-up time,release time and overrun time,and the rule of the variation of the contact resistance is analyzed.In order to avoid the influence of the error of test,wavelet de-noising is used to de-noise the experimental data.At the same time,in order to avoid the problem that the training time is increased due to the singular sample data in the later learning.In this paper,by using the method of the maximum normalization normalized the test data,which provided the basis for establishing prediction model and predicting the storage life of aerospace relay.Then,the applicability of the dynamic neural network in the prediction of storage life of aerospace relay is analyzed.By the analysis of the network structure of nonlinear auto regressive neural network,and the selection of the network mode and model parameters,the life prediction model of aerospace relay based on nonlinear auto regressive neural network is established.The contact resistance is trained,studied and predicted by using this model,and the prediction results are analyzed.Finally,a prediction model of the aerospace relay storage life of support vector machine based on the principle of structural risk minimization is established.Based on the research of the support vector machine prediction model,the selection and optimization of model parameters,the support vector machine life prediction mode of aerospace relay is established.Through the support vector machine prediction model,the contact resistance is trained and predicted,and the prediction results are compared and analyzed.The results show that both prediction models can better predict the storage life of the aerospace relay,and the accuracy of support vector machine prediction model is higher than nonlinear auto regressive neural network neural model.