Reliability Modeling And Burn-in Scheme Optimization Of Degradation Systems

With the increasing market competition,except for the function and performance of a product,its reliability has been regarded as one of the most important evaluating indicators.For manufacturers,improving the reliability of the products and reducing the uncertainty of failure will help them to enhance their brand premium ability,which is an important way to improve the competitiveness of their products and the profits and of their enterprises.Therefore,in recent years,the research of reliability has received extensive attention from industry and academia.With the rapid development of measurement techniques and further understanding of failure mechanisms,there is a growing number of products whose failure can be attributed to their performance degradation,and then degradation-based reliabiltiy analysis has become a hot topic in this area of research.This project comes from the national major scientific equipment and instrument development project "high power semiconductor laser comprehensive test instrument".High power semiconductor laser is the core component of laser industry and a typical device with performance degradation.The output optical power will gradually weaken with the increase of working time,until it can not meet the application requirements,resulting in degradation failure.In this dissertation,based on semiconductor lasers,reliability modeling,residual life prediction,burn-in strategy optimization of degraded systems(products)and the design of reliability test instrument for high power semiconductor laser have been studied.The main contributions of this dissertation are as follows:1.The degradation modeling method based on failure rate is investigated.The increasing failure rate is employed to discribe the system inherent degradation,and random shock is introduced to capture the impact from external environment.And then the system reliability model is derived with the method of effective age,which is used to model the effect between the inherent degradation and cumulative random shocks.Moreover,one maintenance policy which combine preventive maintenance and corrective maintenance is presented,and then a multi-objective maintenance optimization is performed to maximize the system availability,and minimize the system long-fun expect cost,with the two decision variables:period of PM and the number of PMs before replacement.2.The degradation modeling based on inverse gaussian process and the corresponding remain useful lifetime estimation are studied.In this model,the inverse Gaussian process is used to model the degradation paths,random effect parameter is introduced to capture the heterogeneity across individuals in the same population,the normalized covariate is used to represent the working stress,and hence the remain useful lifetime distribution of the degradation process with nonlinear shape function is derived.In the online update stage,the random effect parameter is updated by using Bayesian theory,then the actual working stress could be further estimated and the remaining useful lifetime could be predicted more accurately.This model could take the system working load and the individual heterogeneity into account,and hence improve the prediction accuracy of remaining useful lifetime estimation by updating parameters online.3.The burn-in scheme optimization for degradation products is studied.For those degradataion-failed products,burn-in will degrade their performance and hence diminish their mean lifetime to failure.To address this impact of degaradaion-based burn-in,we develop a multi-objective burn-in method that could simultaneously minimize the burn-in cost and maximize the burnt-in population's mean lifetime to failure.We employ the time-transformed Wiener process with random effects to model the nonlinear degradation path of products and develop a burn-in scheme with two decision variables,namely,test duration and screening cutoff level.Cost expression and lifetime-based optimal objective are analytically developed.The optimal test policy is determined using the multi-objective evolutionary algorithm based on decomposition.Moreover,a simulation study is conducted to demonstrate the usage and effectiveness of this method.4.According to the photoelectric and degradation characteristics of high-power semiconductor lasers,a distributed reliability testing and burn-in screening system based on fieldbus is developed.The design and implementation of the system are discussed in detail from the aspects of the overall structure,hardware circuit design,control algorithm and reliability data analysis of the test system,and its high-efficiency current control and reliability test functions are verified through the tests.The system can be used in the accelerated degradation test,burn-in screening test and life test of high-power semiconductor lasers.