【作者】 吕燚；

【导师】 谢少峰； 章云；

【作者基本信息】 广东工业大学 ， 控制科学与工程， 2019， 博士

【摘要】 随着市场竞争日益激烈,产品除了功能和性能,可靠性也已经成为一项重要的评价指标。对于生产厂家,提升产品可靠性、降低产品失效的不确定性将有助于增强其品牌溢价能力,是提升产品竞争力和企业效益的重要途径。因此,近年来对产品可靠性问题的研究受到了来自学术界和工业界的普遍关注。随着检测技术的进步和对产品失效机理认识的不断深入,越来越多产品的失效可以归因于其性能退化,基于性能退化的可靠性分析已经成为该领域的研究热点之一。本课题来源于国家重大科学设备仪器开发专项“大功率半导体激光器综合测试仪器”。大功率半导体激光器是激光产业的核心部件,是一种典型的性能退化器件。激光器输出光功率随着工作时间的增加会逐渐减弱,直至无法满足应用需求,发生退化失效。本文以半导体激光器为背景,围绕退化系统(产品)的可靠性建模、剩余寿命预测、老练策略优化以及大功率半导体激光器可靠性测试仪器的开发等问题展开研究,主要内容如下:1.研究了基于故障率的退化系统建模与应用。采用故障率递增特性刻画系统自身退化过程,采用随机冲击模型刻画外部环境对系统自身退化过程的影响,基于有效役龄的方法推导出了系统的可靠性模型。提出了一种融合预防性维修和恢复性维修的混合维修策略,并以优化产品长期可用度和平均维修费用为目标对维修策略进行了多目标优化。2.研究了基于逆高斯过程的退化系统建模与剩余寿命预测问题。采用逆高斯过程对系统的退化过程进行建模,引入随机效应参数刻画个体间的退化特性差异,采用双参数环境应力协变量刻画实际环境应力和工作负荷,并推导出了非线性形状函数条件下的剩余寿命分布。根据贝叶斯定理对随机效应参数进行在线更新,实时估计当前工作应力,进而得到当前工作应力下的剩余寿命。该模型在考虑个体差异的同时能够有效地描述实际工作应力对系统固有退化过程的影响,最终通过参数在线更新进一步提高了剩余寿命预测精度。3.研究了具有退化特性产品的老练策略建模与优化问题。对于退化失效的产品,老练试验会导致参与试验的所有产品的性能发生退化,进而影响合格品平均剩余寿命。本文突出了试验过程对产品性能退化的影响,提出了一种最小化老练成本和最大化剩余寿命的老练策略优化模型。采用带有时移特性的维纳过程刻画系统的非线性退化路径,基于退化量构建了以老练试验时间和筛选阈值作为决策变量的老练筛选模型,推导出了试验成本和平均剩余寿命两个维度的目标函数,采用进化算法实现老练策略的双目标优化,并通过仿真实验验证了该方法的有效性。4.依据大功率半导体激光器的光电特性和退化特征,研发了一套基于现场总线的分布式大功率半导体激光器可靠性测试与老练筛选系统。从测试系统的总体结构、硬件电路、控制算法以及可靠性数据分析等方面详细介绍了该系统的设计与实现,并通过测试验证了其高效电流控制与可靠性测试等功能。本系统实现了激光器温度和电流的双应力测试,可用于大功率半导体激光器的加速退化试验、老练筛选试验和寿命测试等多种可靠性试验。更多还原

【Abstract】 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.更多还原