Reliability Analysis Of Software And Hardware Systems Based On Performance Degradation

In the complex system of software and hardware,in addition to the failure of the hardware itself,another major cause is the impact of the use or failure of the computer software.How to effectively diagnose the system and evaluate the reliability of the system has become the current concern research topics.This thesis takes hardware and hardware integrated system as the research object,and conducts in-depth research on the reliability evaluation method of software and hardware integrated system based on its performance degradation process and observation data.The main innovations include the following aspects:First,a fault diagnosis method based on the Mahalanobis distance(MD)area metric.In this thesis,a fault diagnosis method based on Mahalanobis distance metric is proposed.For the monitored historical degradation data and failure data,the multi-dimensional performance degradation data samples are transformed into Mahalanobis distance univariate data by mapping transformation with Mahalanobis distance metric.Then the distribution function is obtained,and the difference between the observed sample and the cumulative distribution function of the univariate data after the training sample is transformed by the area metric method,the area value between the distribution curves is calculated according to the definite integral,and the area size metric is used to diagnose whether the system fails or not,this method converts multi-dimensional indicators into single-dimensional indicators,which is simple to calculate and reduces the complexity and time overhead of multi-dimensional indicator diagnosis.Second,consider the remaining life estimation method of software and hardware integrated complex systems affected by software operation.Based on the hardware performance degradation process,the use or operation of software is regarded as an external impact of the system.A software and hardware system remaining life assessment analysis method considering the impact of software operation is proposed.The method characterizes the impact of software operation on the system through hardware performance degradation index changes,and uses discrete hidden Markov process to describe the relationship between the two.According to the number of change points in the sample of system performance degradation indicators under different software operating conditions,different degradation models are constructed for the same hardware degradation process segmentation,so that the model can describe the degradation process more accurately.Stochastic simulation technology and optimization technology are used to estimate the remaining life of the hardware,and the remaining life of the hardware and software system is estimated according to the system architecture.Third,the residual life analysis model of software and hardware systems based on the ?-Copula model.This chapter proposes a Copula model based on the ?-impact model.According to the monitored degradation data,the biggest advantage of the model is that it does not need to determine the functional relationship between each feature index and the reliability function in advance,because the Copula function for the edge distribution of each indicator is not required,but only the Copula function is needed to obtain the joint distribution of all indicators,and the remaining life prediction value of the system is obtained according to the reliability distribution function.